phenotypic and genomic di erentiation of arabidopsis thaliana...2015/07/22 · 68 swiss alps (luo...

Phenotypic and genomic differentiation of Arabidopsis thaliana1

along altitudinal gradients in the North Italian alps2

Torsten Gunther$,‡,§, Christian Lampei$,‡, Ivan Barilar$, and Karl J. Schmid$,∗3

July 22, 20154

$ Institute of Plant Breeding, Seed Science and Population Genetics, University of Hohenheim,5

Stuttgart, Germany6

§Department of Evolutionary Biology, EBC, Uppsala University, Uppsala, Sweden.7

∗Corresponding author: E-mail: [email protected]

‡These authors contributed equally to this work.9

Abstract10

Altitudinal gradients represent short-range clines of environmental parameters like tem-11

perature, radiation, seasonality and pathogen abundance, which allows to study the foot-12

prints of natural selection in geographically close populations. We investigated phenotypic13

variation for frost resistance and light response in five Arabidopsis thaliana populations rang-14

ing from 580 to 2,350 meters altitude at two different valleys in the North Italian Alps. All15

populations were resequenced as pools and we used a Bayesian method to detect correlations16

between allele frequencies and altitude while accounting for sampling, pooled sequencing and17

the expected amount of shared drift among populations. The among population variation18

to frost resistance was not correlated with altitude. An anthocyanin deficiency causing a19

high leaf mortality was present in the highest population, which may be non-adaptive and20

potentially deleterious phenotypic variation. The genomic analysis revealed that the two21

high-altitude populations are more closely related than the geographically close low-altitude22

populations. A correlation of genetic variation with altitude revealed an enrichment of23

highly differentiated SNPs located in genes that are associated with biological processes24

like response to stress and light. We further identified regions with long blocks of presence25

absence variation suggesting a sweep-like pattern across populations. Our analysis indicate26

a complex interplay of local adaptation and a demographic history that was influenced by27

glaciation cycles and/or rapid seed dispersal by animals or other forces.28

Introduction29

Local adaptation results from interactions of populations with biotic and abiotic environmental30

conditions. Numerous phenotypic traits that contribute to adaptation show therefore clinal vari-31

ation along environmental gradients (Huxley, 1938). Phenotypic clines have been investigated32

for many years (Mayr, 1942) because predictions regarding patterns and processes of natural33

selection can be tested in a rigorous fashion (Etterson & Shaw, 2001). In annual plants, an34

1

not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which wasthis version posted July 22, 2015. ; https://doi.org/10.1101/023051doi: bioRxiv preprint

https://doi.org/10.1101/023051

example of a well studied cline is the change of flowering time with latitude gradients (Stinch-35

combe et al , 2004). The timing of the transition from the vegetative to the reproductive stage36

is strongly related to fitness in annual plants and therefore depends on day length and temper-37

ature.38

In mountaineous environments, altitude is one of the strongest environmental gradients39

because abiotic parameters like temperature, radiation intensity, atmospheric pressure, and40

partial pressure of gases simultaneously change over short distances (Korner, 2007). The biotic41

environment includes competitors, herbivores, pathogens or symbionts whose occurrence differs42

with altitude as well. The combination of these factors results in different ecological niches at43

different altitudes, which likely exert a strong selection pressure for local adaptation. In addition44

to external factors, the intrinsic genetic potential for adaptation is affected by altitude since45

the absolute area available for plant growth tends to rapidly decrease with altitude (Korner,46

2007) causing small population sizes and increasing levels of genetic drift and inbreeding at high47

altitudes (Thiel-Egenter et al , 2011; Manel et al , 2012).48

Another critical factor in adaptation to different altitudes is the age of local populations.49

Long-term changes in temperature and other environmental parameters caused by glacial cy-50

cles contributed to large fluctuations in the altitudinal distribution of species, which retreated51

from high altitudes during glacial maxima and expanded during interglacial periods. However,52

numerous stable glacial refugial populations on mountain tops (nunataks) existed in the Euro-53

pean Alps (Schonswetter et al , 2005) and other mountain ranges. In such populations, selection54

and genetic drift likely contributed to the rapid genetic and phenotypic divergence from other55

populations of the same species, and frequently led to the formation of endemic sibling species56

restricted to high altitudes.57

Arabidopsis thaliana shows a wide distribution and local adaptation within the Northern58

hemisphere for many morphological, life history and other fitness-related traits (reviewed by59

Koornneef et al , 2004; Bergelson & Roux, 2010). Genomic surveys suggested adaptation to60

large-scale environmental gradients in central Europe (Clark et al , 2007; Hancock et al , 2011;61

Horton et al , 2012), the Iberian Peninsula (Mendez-Vigo et al , 2011) and Scandinavia (Long62

et al , 2013; Huber et al , 2014). Adaptation to different altitudes can also be studied in A.63

thaliana since it occurs from the sea level up to 4,250 m altitude (Al-Shehbaz & O’Kane, 2002).64

High altitude adaptation was observed in populations in the eastern Pyrenees and included65

traits like fecundity, phenology and biomass allocation, suggesting selection for higher vigour66

in high altitude plants (Montesinos-Navarro et al , 2011), selection for alpine dwarfism in the67

Swiss Alps (Luo et al , 2015), or a constitutive protection against UV-B radiation protection in68

Shakdara, Central Asia at 3,063 m a.s.l (Biswas & Jansen, 2012).69

We investigated the roles of adaptive evolution and demographic history in the distribution70

of genetic variation and putative adaptive phenotypic traits along an altitudinal gradient in the71

European Alps. We sampled five A. thaliana populations at different altitudes up to 2,300 m72

in the South Tyrolia and Trento provinces of Italy. Four populations represent two pairs of73

high and low altitudes in two different valleys, and and a fifth population was located at equal74

2


https://doi.org/10.1101/023051

distance to both gradients. In these populations, we characterized phenotypic differentiation in75

frost resistance and response to light and UV-B stress because these environmental parameters76

are strongly associated with altitude (Blumthaler et al , 1997; Korner, 2007). Accessions from77

the Northern Italian Alps were previously shown to be genetically distinct from other regions78

and to harbour reduced genetic diversity (Cao et al , 2011). We reasoned that high altitude79

adaptation resulted in a strong signal of genetic differentiation relative to the genome-wide80

background, which should allow to identify selected genomic regions. We sequenced pools of81

individuals from each of the five populations, because pool sequencing is a cost-efficient means to82

analyse diversity in populations to infer demography and selection, and was previously applied83

in several plant and animal species (e.g. Turner et al , 2010; He et al , 2011; Kolaczkowski et al ,84

2011; Fabian et al , 2012; Lamichhaney et al , 2012; Orozco-terWengel et al , 2012; Fischer et al ,85

2013). Although sequencing of pools comes at the cost of noisy estimates and a loss of linkage86

information (Futschik & Schlotterer, 2010; Zhu et al , 2012), suitable frameworks that account for87

the special properties of pooled sequence data are available (Futschik & Schlotterer, 2010; Kofler88

et al , 2011b,a; Boitard et al , 2012; Gunther & Coop, 2013). We investigated the demographic89

history by Approximate Bayesian Computation (ABC, Beaumont et al , 2002), and correlated90

allele frequencies with altitude to identify genomic regions that are significantly differentiated91

between high and low altitude populations because of local adaptation. We observed a complex92

pattern of genomic and phenotypic diversity suggesting that adaptation to high altitudes may93

affect genes involved in light responses and soil conditions as well as other stress factors, but94

also indicate that there is no simple pattern of genomic and phenotypic differentiation with95

altitude.96

Materials and Methods97

Plant material98

Seeds were sampled at five locations in the alps of South Tyrolia and Trento provinces to99

represent two pairs of high and low altitudes, respectively, in different valleys and mountain100

ranges but in close geographic proximity (Table 1 and Figure 1). All populations were located101

on steep slopes and exposed to the South (Coordinates: Vioz/Coro 46°22’57”N, 10°39’35”E;102

Terz 46°21’50”N, 10°55’13”E; Finail 46°44’31”N, 10°49’5”E; Juval 46°38’55”N, 10°58’27.85”E;103

Laatsch 46°40’18.00”N, 10°30’54.00”E). Two low altitude populations (Laatsch and Juval) are104

located on scree which was partly overgrown in Juval, but still moving in Laatsch. These two105

sites were the least disturbed of the five sites. The third valley population (Terz) is located106

on marginal soils in dry stone walls in the vicinity of a wine yard and was highly disturbed.107

The two high-altitude populations occur at South exposed rocky outcrops that serve as resting108

places of mountain goats, and therefore are strongly disturbed habitats with a high supply of109

nitrogen and phosphorous. Due to different seasonality, we visited the low altitude populations110

in early May and the high altitude populations in mid-July. All locations were sampled in111

a spatially balanced design to include the maximal diversity present at the site. Plants were112

3


https://doi.org/10.1101/023051

Table 1: Summary of pooled sequencing libraries.

PoolSeq summary statistics

Individuals Median coverage Percent NucleotidePopulation sequenced Altitude [m] of nDNA genome covered diversity, π (SE) Tajima’s D (SE)

Juval 29 577 - 657 32x 94.3 0.00305 (0.00003) 0.0352 (0.0063)Terz 25 810 - 945 31x 95.8 0.00439 (0.00003) -0.0107 (0.004)Laatsch 29 970 - 1010 19x 94.5 0.00190 (0.00003) -0.1912 (0.0084)Finail 29 2250 - 2270 32x 94.2 0.00190 (0.00002) -0.7567 (0.0046)Vioz/Coro 19 2210 - 2355 50x 93.3 0.00166 (0.00002) -0.7122 (0.0044)

cultivated for DNA extraction on standard gardening soil (Einheitserde ED 63T). Forty day old113

whole plants were freeze dried in a Christ Alpha 14 (Martin Christ Gefriertrocknungsanlagen114

GmbH, Osterode am Harz, Germany) freeze dryer before DNA extraction.115

Phenotypic analysis of spring freezing tolerance and response to light stress116

and UV-B radiation117

Freezing tolerance We investigated the plants for differences in freezing damage on the118

population level under two temperature regimes with plants from all populations grown in119

the greenhouse by adapting the method of Martin et al (2010). Eight random accessions per120

population were tested at a minimum temperature of -15°C, and five random accessions per121

population at -7°C minimum temperature. We raised the plants under long day condition (16122

h light/8 h dark) at 23° C and adapted them to the cold for one day at 4° C at the eight leaf123

stage. Whole plants were sampled, washed with deionized water and wrapped in aluminum foil124

after removal of residual water.125

We then put the samples into a steel vacuum Dewar bottle which was placed in a styrofoam126

box (14 l volume and 5 cm wall thickness). Two temperature loggers (3M�Temperature Logger127

TL20, 3M Deutschland GmbH, Neuss) were added to the Dewar bottle above and below the128

samples to monitor the realized treatment temperature. To ensure a slow temperature decrease,129

we added a PET-bottle with 1 l of saturated salt water (20% NaCl solution) to the styrofoam130

box. The box was kept at -20° C for 17 and 7 hours, respectively. Slow defrosting was ensured131

at 4°C for approximately 20 h. Samples were put into 10 ml of deionized water in sterile 15 ml132

centrifuge tubes and left at room temperature overnight to allow cytosol leakage of damaged133

cells. We then measured electric conductivity with a handheld conductivity meter (WTW LF134

90, KLE1 sensor; measure 1). To obtain conductivity at 100% cell damage, we autoclaved135

all samples (Systec DB-23 benchtop autoclave) and measured conductivity again (measure 2).136

Freezing damage of each sample was calculated as the proportion of measure 1 over measure 2.137

Light and UV-B stress We further characterized the response to different light conditions.138

Random individuals of each population were cultivated on standard gardening soil (Einheitserde139

ED 63T) at 23� in a climate cabinet (GroBank BB-XXL3+, CLF Plant Climatics, Emersacker,140

Germany) under three light treatments under long day conditions (16/8 light/dark): (1) A141

4


https://doi.org/10.1101/023051

JuvalLaatsch

Finail

TerzVioz/Coro

LaatschHigh altitude populationsLow altitude populations

10°18' 10°28' 10°38' 10°48' 10°58' 11°08' 11°18' 11°28'

46°0

8'46

°16'

46°2

4'46

°32'

46°4

0'46

°48'

Longitude

Latit

ude

Figure 1: Geographic map showing the location of the five populations included in this study.Map source: http://www.maps-for-free.com/

control low light treatment, (2) a low light with UV-B treatment and (3) a light-stress treat-142

ment without UV-B. The low light treatment consisted of 168 µmol m−2 s−1 fluorescent light143

(Philips Master TL-D 58W/840 Reflex), and the UV-B treatment used the same fluorescent144

setting with additional UV-B radiation exposure three times per day for one hour at variable145

intensities (6:00−7:00 am 0.01 mWcm−2, 11:00−12:00 am 0.027 mWcm−2 and 4:00-5:00 pm146

0.027 mWcm−2). In the light-stress treatment, we exposed plants to fluorescent light with an147

intensity of 470 µmol m−2 s−1, which is more than twice the optimal light intensity for the Col-0148

ecotype (200 µmol m−2 s−1). In the UV-B treatment, fluctuating UV-B levels were judged to149

be more realistic than stable UV-B exposure because frequent cloud cover reduces the total150

annual amount of light and UV-B radiation but fluctuations in light intensity increase strongly151

with higher altitude (Korner, 2007). The leaf color of each plant was used as stress indicator152

because light or UV-B stressed plants show increased anthocyanin production which causes a153

red leaf colour (Chalker-Scott, 1999). All plants were photographed with a digital camera (Sam-154

sung NX 10) at high resolution from the top and customized macros of the ImageJ distribution155

Fiji (Schindelin et al , 2012) were used to measure the red leaf area, the total leaf area and the156

5


http://www.maps-for-free.com/

https://doi.org/10.1101/023051

area of dead plant material.157

Statistical analysis of phenotypic data To identify differences in freezing damage between158

populations, we fitted a generalized linear model (GLM) with freezing damage odds as dependent159

variable and a population factor and a test-condition factor and their interaction as independent160

variables based on a quasibinomial error distribution (logit link) as implemented in the R161

package stats (DevelopmentCoreTeam, 2014). Although statistical theory suggests to use a162

binomial error distribution to analyse the proportion of read and dead tissue after the light163

treatment, we used a general least squares model (GLS) assuming a normal distributed error164

because the residual distribution did not deviate from a normal distribution (Shapiro-Wilk test,165

W = 0.96, p = 0.19) and only the GLS model allows to correct for the strong heteroscedasticity166

in the data. Using the nlme package in R (Pinheiro et al , 2014) we fitted GLS models with the167

red proportion or the dead proportion as dependent variable and population, treatment and168

their interaction as independent factors. We allowed for different variances for each level of169

population and treatment using the varComb function with two varIdent functions. To identify170

differences between treatments in both experiments, we performed post-hoc comparisons with171

the glht function from the R package multcomp and corrected for the false discovery rate (FDR;172

Benjamini & Hochberg, 1995).173

Sequencing and population genetic analysis174

DNA sequencing We ground dry whole plants with a Retsch MM400 centrifugal mill (Retsch175

GmbH, Haan, Germany) and extracted the DNA with a standard CTAB maxi prep protocol.176

DNA quality and concentration were determined on a 0.8% agarose gel and additionally with177

the Qubit 2.0 Fluorometer. For each of the five populations, we prepared a pooled DNA sample178

(5 µg) by mixing of equimolar amounts of DNA from all individuals. Each DNA pool was179

then converted into a barcode-tagged genomic sequencing library using the standard Illumina180

protocol. The five tagged libraries were pooled and sequenced together on a single lane of an181

Illumina HiSeq 2000 sequencer by GATC Biotech (Konstanz, Germany). We then trimmed182

the 101 bp long paired-end reads to consecutive runs of base calls with a quality ≥ 25 and183

filtered for a minimum read length of 35 bp using SolexaQA (Cox et al , 2010). The trimmed184

high-quality reads were mapped against the A. thaliana TAIR10 reference genome (Lamesch185

et al , 2012) using BWA version 0.6.2 (Li & Durbin, 2009) using the non-default parameters186

-n 0.05 (mismatch rate) and disabled seeding of reads. The BWA module sampe mapped187

paired-end reads to a single SAM file and samtools version 0.1.18 (Li et al , 2009) filtered the188

resulting alignments to include only properly paired reads and a minimum mapping quality of189

20. For all downstream analyses, we converted the data to pileup files with samtools mpileup.190

We then identified SNPs as polymorphic sites of at least 12x coverage in at least one pool with191

a minimum of two reads supporting a non-reference allele and polarized them with the close192

relative Arabidopsis lyrata (Hu et al , 2011).193

6


https://doi.org/10.1101/023051

Population genomic analysis We further analysed the Pileup files with Popoolation version194

1.2.2 (Kofler et al , 2011a) to calculate nucleotide diversity, π, Wattersons θ and Tajima’s D195

separately for each population. First, each position was independently subsampled to create196

a uniform 18x coverage over all populations by random drawing base calls from the pileup file197

without replacement. Then, π, θ and Tajima’s D were calculated separately for each population198

using sliding windows with window and step sizes of 2,000 bp each. Only sites with a coverage199

of at least 10 and maximally 100 were included in the next analysis steps.200

We annotated SNPs with SnpEff version 3.1 (Cingolani et al , 2012) based on the TAIR10201

annotation (Lamesch et al , 2012) and assigned them to one of the following classes: intergenic,202

5’ UTR, 3’ UTR, intron, synonymous coding, non-synonymous coding, synonymous stop, stop203

lost, stop gained and start lost. SNPs belonging to different groups (e.g. non-synonymous204

coding and stop gained) were assigned to the stronger effect category (e.g., stop gained). We205

further annotated regulatory SNPs based on known and predicted transcription factor binding206

sites from AGRIS (Yilmaz et al , 2011). The effect of non-synonymous SNPs on protein function207

(Gunther & Schmid, 2010) was determined with the MAPP program (Stone & Sidow, 2005),208

which classifies nonsynonymous SNPs into deleterious or tolerated amino acid substitutions209

based on a collection of homologous proteins obtained with PSI-BLAST searches (Altschul210

et al , 1997) of all A. thaliana proteins against Uniprot (Wu et al , 2006). We calculated a211

phylogenetic tree for these proteins with semphy (Friedman et al , 2002) and used the multiple212

protein sequence alignment and the tree as input for MAPP (Stone & Sidow, 2005).213

Identification of strongly differentiated genes To identify strongly differentiated genes,214

we correlated altitude with allele frequencies within populations with Bayenv2.0 because it215

models the sampling noise inherent to pooled sequencing and accounts for a shared history of216

populations (Gunther & Coop, 2013). To characterize this relationship, we estimated a covari-217

ance matrix with Bayenv2.0 from a subsample of 10,000 high confidence SNPs (i.e. SNPs with218

a coverage between 30 and 100 in each of the five populations) based on 1,000,000 MCMC iter-219

ations. To ensure convergence we estimated a second matrix from another subsample of 10,000220

SNPs. For the Bayenv2.0 analysis, we included only SNPs for which both alleles segregated in at221

least two populations and the per population sequence coverage was between 15 and 100, which222

resulted in a total of 400,231 SNPs. Using these data we estimated the genotype-environment223

correlation statistics Z and ρ(X,Y ′) (Gunther & Coop, 2013) between allele frequencies and224

altitude per population (normalized to mean 0 and standard deviation 1) estimated during225

200,000 MCMC iterations.226

To further analyse the functional annotation of highly differentiated genes, we searched227

for an enrichment of different annotation categories obtained from AraPath (Lai et al , 2012),228

which comprises gene ontology (GO, Ashburner et al , 2000), KEGG (Kanehisa et al , 2012),229

AraCyc (Mueller et al , 2003) and plant ontology (PO, Cooper et al , 2013) annotations for all A.230

thaliana genes, and additionally includes gene sets obtained from extensive literature research231

(Lai et al , 2012). To avoid a biased enrichment analysis, we tested enrichment of annotations232

using Gowinda (Kofler et al , 2012), which accounts for gene length using a permutation ap-233

7


https://doi.org/10.1101/023051

proach. We conducted 100,000 permutations and included all annotation categories of at least234

two genes while all SNPs in a gene and within 2,000 bp distance of a transcript were assigned235

to the respective gene.236

Demographic history We investigated the genetic relationship of populations with Neighbour-237

joining trees (Saitou & Nei, 1987) based on the matrix of average pairwise FST values calculated238

by PoPoolation2, and from a correlation matrix of allele frequencies derived from the covariance239

matrix estimated by Bayenv2.0. The trees were plotted with the R package ape (Paradis et al ,240

2004). Additionally, we reconstruced the historical relationship among the five populations241

using their current genome-wide allele frequencies with TreeMix (Pickrell & Pritchard, 2012).242

We carried out a run which included all non-singleton SNPs segregating in all five populations243

with a minimum coverage of 20 per population (1,071,527 SNPs). We bootstrapped each run244

1,000 times and carried out a four-population test which tests whether the relationship of the245

two pairs of high and low populations can be explained by any of the possible tree topologies246

(Reich et al , 2009; Patterson et al , 2012) to verify the TreeMix results.247

We estimated divergence times between pairs of high and low altitude populations by sim-248

ulating models of demographic history with msABC (Pavlidis et al , 2010) followed by an Ap-249

proximate Bayesian Computation (ABC) analysis with the abc R package (Csillery et al , 2012).250

We investigated four different models under a population-split scenario to estimate divergence251

times between low and high altitude populations (Figure S3). All models assumed a common252

ancestral population with a founding bottleneck leading to the high altitude population, but253

models differed in their post-split growth parameters of the high altitude populations: (i) the254

no-growth model assumed a constant population size after the split, (ii) the step-growth model255

assumed an instantaneous population size change some time after the split, (iii) the growth256

model assumed an exponential growth of the population after the split where the growth rate257

was calculated based on a previously drawn current population size (N0; Table S3) and the258

time of the split, and (iv) the exponential model assumed exponential growth of the population259

size after the split. We simulated each model 200,000 times with and without bidirectional260

migration and included them in a model selection step with cross-validation based on selected261

summary statistics (see below) to select the best-fitting model. Posterior model probabilities262

were calculated with the postpr function of the R abc package using the mnlogistic method and263

a tolerance rate of 0.005. Using the best fitting model, we inferred the posterior distribution of264

the population split time.265

In the simulations we assumed uniform priors for parameters (Table S3), and obtained266

mutation and recombination rates used in simulations from empirical studies (Kim et al , 2007;267

Ossowski et al , 2010; Cao et al , 2011) or based them on plausibility arguments (size of the268

high altitude populations, population growth parameter, time of the split). The ABC analysis269

included the following summary statistics: mean and variance of Tajima’s D and π, mean of270

Watterson’s θ and average number of segregating sites for each of the populations; FST values271

and numbers of shared, fixed and private SNPs, which were calculated for each population pair.272

We calculated summary statistics from 2 kb windows and randomly sampled 1,000 windows from273

8


https://doi.org/10.1101/023051

the sequence data. Posterior parameter distributions were estimated by comparing observed and274

simulated summary statistics using the abc function of the R abc package with the rejection275

method and a tolerance rate of 0.01. From the posterior distributions of the best models,276

we calculated the Highest Posterior Density Interval (HPDI 90) of the estimated mean of the277

population sizes and the time of their split using the LaplacesDemon R package (Hall, 2013).278

The HPDI 90 measurement gives the most accurate 90% credible interval (Bayesian equivalent279

of confidence intervals) of the posterior distributions.280

Results281

Phenotypic differences between high and low altitude populations In the test for282

freezing resistance we first decreased the temperature to -15°C with a shoulder at around -7°C,283

which marks the begin of major tissue freezing (Figure 2 B). In the second experiment, tem-284

peratures were lowered to only -7° C and no shoulder was observed in the temperature curves285

(Figure 2 D). The two freezing treatments caused strong differences in the degree of tissue dam-286

age (Figure 2 A+C, F1/59=128, p<0.001) with a population median from 0.8 to 0.9 in the first287

test and from 0.1 to 0.6 in the second test. Based on the GLM there was no significant tempera-288

ture × population interaction (F4/55=1.42, p=0.24), but differences in freezing damage between289

populations (F4/59=7.16, p<0.001). A post-hoc multiple comparison test (α=0.05, FDR cor-290

rected following (Benjamini & Hochberg, 1995)) combined the populations Juval, Laatsch and291

Vioz/Coro into a homogeneous subset with stronger leaf damage and the populations Terz and292

Finail in a subset with higher frost resistance (Figure 2). The differentiation suggests adap-293

tation to local temperature regimes, but since both subsets include at least one high and one294

low altitude population and originated from different drainage systems, frost resistance in these295

populations is not strongly correlated with altitude nor with physico-geographic proximity.296

As second phenotypic trait we investigated response to different types and strengths of light297

irradiation measured as red coloration of leaves, which is caused by anthocyanin accumulation298

and represents a typical stress reaction of plants (Park et al , 2007). The population × treatment299

(light condition) interaction was highly significant (F8/28=42, p<0.001) indicating differential300

population response to light stress. A post-hoc multiple comparison test (α=0.01, FDR cor-301

rected following Benjamini & Hochberg (1995)) on the GLS model indicated that plants from302

four of the five populations developed more red tissue under light stress than in the other light303

conditions (Figure 3 A). Conversely, plants of the highest population (Vioz/Coro) were green in304

all treatments (Figure 3 A, D) but developed more dead tissue under the high light treatment305

in comparison to the low light and UV-B treatments (Figure 3 B, D, population × treatment306

interaction F8/28=14, p<0.001).307

An increased sensitivity to UV-B damage is expected in anthocyanin-deficient plants if the308

flavonoid pathway is affected simultaneously (Chalker-Scott, 1999). To test whether plants309

from high-altitude Vioz/Coro site showed an increased UV-B sensitivity, we exposed them to a310

gradient of increasing UV-B dosages over a period of three days and evaluated the proportion of311

dead tissue (Supplementary Methods). In contrast to the Finail population, which exhibited a312

9


https://doi.org/10.1101/023051

●

600 850 1000 2260 23550.0

0.2

0.4

0.6

0.8

1.0

prop

ortio

n da

mag

ed ti

ssue

low high

A

group Agroup B

−15° Cn=8

0 5 10 15 20 25 30 35

−15

−10

−5

0

5

10

15

time (h)

tem

pera

ture

(C

°)

B

●

600 850 1000 2260 23550.0

0.2

0.4

0.6

0.8

1.0

altitude

prop

ortio

n da

mag

ed ti

ssue

Juval Terz Laatsch Finail Vioz/Coro

low highC −7° Cn=5

0 5 10 15 20 25 30 35−15

−10

−5

0

5

10

15

time (h)

tem

pera

ture

(C

°)D

Lowest temperature − 15° C

Lowest temperature − 7° C

Figure 2: Freezing damage measured as cytosol leakage of destroyed tissue after freezing down to-15° C (A) and -7° C (C) as a fraction of cytosol leakage from completely destroyed tissue (n = 8).The vertical line divides the figure into the low altitude and the high altitude populations. GroupA and B indicate homogeneous population subsets (α = 0.05) after correction for false discoveryrate. (B) and (D) show the logged temperature as area between the measurements of two dataloggers positioned above and below the probes.

superior UV-B resistance compared to the low altitude populations, plants from the Vioz/Coro313

population developed an increased proportion of dead tissue with higher UV-B dosage (binomial314

GLM; population × UV-B dosage F4/186=22, p<0.001; Figure S1, Table S1).315

10


https://doi.org/10.1101/023051

Figure 3: Phenotypic effects of 20 days of light stress treatment. A) Proportion of red tissuein light-stressed plants along the altitude gradient. B) Proportion of dead leaf tissue in light-stressed plants along the altitude gradient. Significant treatment differences at p < 0.001 orsmaller (corrected for false discovery rate) are indicated with a star. C) A representative plantfor the light stress treatment from the Finail population. D) A representative plant for the lightstress treatment from the Vioz/Coro population.

Genetic diversity within and genetic relationship between populations Pool sequenc-316

ing of all five populations and subsequent mapping to the Col-0 genome resulted in a 19- to317

50-fold median coverage of the nuclear chromosomes per population (Table 1). In each pool,318

about 95% of the reference genome sequence positions were sequenced at least once and a total319

of 3,075,350 SNPs were called. We first calculated nucleotide diversity, π, and Tajima’s D by320

accounting for pooling (Futschik & Schlotterer, 2010; Kofler et al , 2011a). As observed before321

(Clark et al , 2007; Cao et al , 2011), diversity was elevated in pericentromeric regions relative to322

chromosome arms (Figure 4A). The genome-wide average diversity of low altitude populations323

was high, while average Tajima’s D was close to 0 (Figure 4B, Table 1). In contrast, high324

altitude populations show a reduced nucleotide diversity and negative Tajima’s D values.325

Allele frequency distributions were similar across populations with some interesting differ-326

ences (Figure S4). All populations showed a large number of rare alleles and a substantial327

number of high-frequency or fixed derived alleles. The proportion of SNPs segregating at in-328

termediate frequencies differed between populations (Table 1, Figure S4). SNPs with a strong329

11


https://doi.org/10.1101/023051

JuvalLaatsch

FinailTerz

Vioz−Coro0

12

Nuc

leot

ide

dive

rsity

π x

10−2

A

0 5 10 15 20 25 30

−1.

5−

1−

0.5

00.

51

1.5

Mbp

Tajim

a's

D

B

Figure 4: Sliding window analysis of (A) nucleotide diversity, π and (B) Tajima’s D for chro-mosome 1 in all five populations.

putative functional effect (annotated as non-synonymous coding, stop lost, stop gained, start330

lost) segregate at much lower allele frequencies than other SNP types (Wilcoxon tests, p < 10−7331

for all classes when compared to synonymous SNPs in each population). Allele frequencies of332

deleterious nsSNPs (as classified by MAPP) were lower than of tolerated nsSNPs in all pop-333

ulations (Wilcoxon tests, p < 10−5), indicating that purifying selection acted on this class of334

alleles. The ratio of nsSNPs to sSNPs, however, did not differ between high and low altitude335

populations (Fisher’s exact test, p > 0.05) suggesting that the strength of selection against336

deleterious polymorphisms was similar in high and low altitude populations.337

The pairwise correlations of allele frequencies derived from the covariance matrix estimated338

by Bayenv2.0 and pairwise FST values (Table S2) revealed that both matrices were correlated339

(Mantel-test, r = −0.63, p = 0.035). Pairwise FST values were similar between all pairs of340

populations, although the maximal distance was observed between Vioz/Coro and Terz popula-341

tions, which belong to the same drainage system. This pair also showed the lowest correlation of342

allele frequencies as estimated by Bayenv2.0. Generally, the correlation estimates gave a higher343

variation than the FST values across population pairs (Table S2), because Bayenv2.0 accounts344

12


https://doi.org/10.1101/023051

for different sequence coverage in the correlation matrix.345

Neighbour-Joining trees generated from the two distance matrices indicate a closer relation-346

ship between the two high altitude populations than to the low altitude population from the347

same drainage system (Figure 5). The FST -based tree is almost star-like with short internal348

branches, whereas the tree derived from the covariance matrix shows longer internal branches349

and a stronger clustering of population pairs. We observed essentially the same clustering with350

a TreeMix analysis based on 1.07 million SNPs with a minimum coverage of 20 per population351

(Figure 5C). Bootstrapping results of the TreeMix analysis show a 77% confidence for the Juval352

and Terz population group and a migration event with 73% confidence from the Finail to the353

Juval population. The four-population test (Reich et al , 2009) rejected all possible tree topolo-354

gies (all |z| > 3, which corresponds to p < 0.001), but the lowest |z| was observed for the tree355

that groups the two high altitude populations together (Table 2). The positive test statistic for356

this tree suggests some admixture in at least one drainage system, most likely between Finail357

and Juval, consistent with the TreeMix results.358

Table 2: Four-population test of two high and two low-altitude populations. It tests the prob-ability that the relationship of four populations can be explained by one of three possible treetopologies without migration. The test was conducted with 1.95 Million SNPs with a minimumcoverage of 20 in each population.

Population pairs f4 statistic Std. error z-score

(Juval,Finail) vs. (Terz,Vioz/Coro) 0.0160 0.0019 8.28(Juval,Terz) vs. (Finail,Vioz/Coro) 0.0032 0.0010 3.10(Juval,Vioz/Coro) vs. (Finail,Terz) -0.0128 0.0020 -6.38

Modelling the demographic history with ABC The tree-based analyses suggested a359

common origin of the two high altitude populations. We therefore considered each pair of360

low and high altitude populations from the same drainage system as an independent replicate361

of the same demographic process to estimate divergence time between high and low altitude362

populations. In both analyses the exponential growth model with bidirectional migration gave363

posterior probabilities of almost 100% (Table S4) with very similar parameter estimates. The364

effective population size of the high altitude populations was smaller than of the low altitude365

populations after the population split (Table 3), with up to a three-fold difference in population366

size. The 90% HPDI of the split time in the models ranged from 14,800 to 260,000 years before367

present, and the interval was almost identical between the two population pairs (Table 3).368

Highly differentiated SNPs between high and low altitude populations To identify369

candidate loci for altitude specific adaptation, we correlated allele frequencies of populations370

with altitude using Bayenv2.0. The Z (Figure S5) statistic (Gunther & Coop, 2013) was calcu-371

lated for 400,231 SNPs in total after quality control (see methods section). A total of 25 SNPs372

showed the highest score possible for Z corresponding to a strong support for a non-zero corre-373

lation (i.e. Z = 0.5; Table S6). One non-synonymous SNP among them was found in the gene374

13


https://doi.org/10.1101/023051

Figure 5: Relationship between populations based on measures of pairwise genetic differentiationand on allele frequencies. Branch lengths reflect the extent of genetic differentiation. (A) NJtree of the correlation matrix obtained with Bayenv2.0. (B) NJ tree of the matrix of pairwiseFST values. (C) Tree inferred with TreeMix of the five alpine populations with migration eventsbased on 1.07 Million SNPs with a coverage of > 20. The migration event is coloured accordingto its weight which represents the percentage of alleles originating from the source. Numbersat branching points and on the migration arrow represent bootstrapping results based on 1,000runs.

AT5G17860, which codes for the Calcium exchanger 7 (CAX7) protein located on chromosome375

5 (Figure 6A, B). A region about 5 kb upstream of this gene showed no sequence coverage in376

both high altitude populations and the Laatsch population (Figure 6B, C). This pattern indi-377

cates that the Col-0 reference haplotype is not present in these populations, either because of a378

deletion or the presence of a highly divergent haplotype of a length of about 25 kb. The whole379

region was flanked by SNPs with high Z values and strong differentiation between high and380

low altitudes (Figure 6A). To test whether this pattern resulted from insufficient local sequence381

coverage, we analysed all 33,268 TAIR10 genes with sequence data in at least one population.382

Only 15 genes showed a similar pattern, which we defined as a per base coverage below the383

5th percentile in high altitude and above the 5th percentile in low altitude populations. This384

14


https://doi.org/10.1101/023051

Table 3: Posterior probability of the parameters of the best-fitting demographic model (expo-nential growth, with migration). The probability for each parameter was inferred by ABC.

Juval - Finail Terz - Vioz/Coro

Parameter Mode 90% HPDI Mode 90% HPDI

Exponential growth, with migration Exponential growth, with migrationNlow alt. 195,457 168,005 - 199,983 196,573 176,779 - 199,988Nhigh alt. 71,221 19,460 - 166,695 77,039 18,275 - 169,580Tsplit 70,650 18,160 - 261,636 77,491 14,892 - 260,579

low probability to find similar presence-absence variation in our sequence data supports a dele-385

tion in the high altitude populations. Of these 15 genes, 12 genes were annotated as unknown386

protein, protein of unknown function, pseudogene or transposon; one gene encodes a ribosomal387

protein, one gene a cytochrome P450 family protein and the last one is an NBS-LRR receptor388

(AT5G48770).389

A second candidate gene encodes the Eps15 homology domain protein AtEHD1 (AT3G20290)390

where a single SNP located in an intron shows a high Z-score (Z = 0.5) but is flanked in a region391

of at least 50 kb by additional SNPs highly associated with altitudes. This region includes two392

additional genes downstream of the AtEHD1 gene (Figure S7). In general the proportion of393

fixed alleles in the high-altitude populations is much larger than in the low-altitude popula-394

tions. Here, only a single SNP shows the highest correlation with altitude, however, across a395

large region the two high-altitude populations were strongly differentiated from the other two396

populations.397

A third example of an outlier region is located on chromosome 3 around a highly differenti-398

ated SNP in the gene AT3G07130 which encodes the purple acid phosphatase 15 (Figure S8).399

In this region, only the Vioz/Coro population at the highest altitude has a large number of400

fixed alleles and is strongly differentiated from the other four populations, which is consistent401

with a footprint of selective fixation of alleles in this population.402

Enrichment analysis of highly differentiated genes To test whether functional groups403

of genes show a higher level of differentiation, we considered the top 1% of SNPs as highly404

differentiated SNPs and conducted an enrichment analysis with Gowinda. This analysis did not405

uncover enriched gene sets after correction for multiple testing with the exception of gene sets406

obtained from a literature collection (Lai et al , 2012) (Table S11). Nevertheless, categories with407

nominal p-values ≤ 0.05 included annotations expected to be involved in adaptation to different408

altitudes (Tables S7 and S8). For example, biological processes like response to light stimulus,409

heat acclimation, defence signalling pathway, flavonol biosynthetic process and biochemical410

processes like superoxide radical degradation, glucosinolate biosynthesis, flavone and flavonol411

biosynthesis and anthocyanin biosynthesis are some of the processes which are likely involved in412

adaptation to altitude. We also annotated the functional effects of highly differentiated SNPs413

(Figure S6). Genes highly differentiated between high and low populations were enriched for414

15


https://doi.org/10.1101/023051

genic SNPs because most SNPs were located in introns and UTRs, but not in coding regions.415

If genic but non-coding SNPs were involved in adaptive differentiation, selection should have416

preferentially acted on regulatory processes rather than on protein function. Such a hypothesis417

was supported by an enrichment analysis of highly differentiated SNPs in known predicted418

transcription factor binding (TFB) sites. According to the AGRIS database (Yilmaz et al ,419

2011) 3.3% of all SNPs were located in TFB sites, but highly differentiated SNPs were enriched420

in TFB sites although effect sizes were small (Bayenv2.0 candidates: 3.75%; Fisher’s exact test,421

p = 0.03711).422

16


https://doi.org/10.1101/023051

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●● ●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●● ●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●● ●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●● ●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●● ●●●●● ●●●●●●●●●●●●●●●●●●●● ●●●●●

●●●●●

●●●●●●●●●● ●●●●● ●●●●●

●●●●●

●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●● ●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

Position [bp]

geno

me−

wid

e ra

nk (

Z)

●●●●●

0.0

0.2

0.4

0.6

0.8

1.0

5880000 5890000 5900000 5910000 5920000 5930000

AAT5G17780

AT5G17790

AT5G17795AT5G17800

AT5G17810

AT5G17820AT5G17830

AT5G17840

AT5G17847AT5G17850

AT5G17860

AT5G17870AT5G17880

AT5G17890

AT5G17900AT5G17910

AT5G17920

AT5G17930

Focal gene NB−LRR gene other gene

Position [bp]

Alti

tude

[m]

5880000 5890000 5900000 5910000 5920000 5930000

617

877

990

2260

2282

B ●

●

●

●

●

●

●INTERGENICNON_SYNONYMOUS_CODING

UTR_5_PRIMESYNONYMOUS_CODING

INTRONUTR_3_PRIME

STOP_GAINED

●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●● ●● ●●●● ● ●● ●●●●● ●●●●●●●●●●●●●●●●● ● ●●● ●●●● ● ●● ●●●●●● ●● ●● ●●●●●●● ● ●● ●●●●●●●●● ●●●● ●●●● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●● ●●●●●●●●●●●●●●●●●●● ●●●●●●● ●●●●●●●●●●●●● ●●● ●●●●●●●●● ● ●●● ●●● ●●●●●●●● ● ● ●● ● ● ●● ● ●● ●● ● ●● ●●●●●●● ●●● ● ●●● ●●●●●●●●●●●●●● ● ● ●●● ●● ●●●●●● ●● ● ●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●● ●●●●●●●●● ● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●● ●●●●●●●● ●●●● ●●●●●●●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●● ●●●●● ●● ●●●● ●●●●● ●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●

Position [bp]

Cov

erag

e pe

r po

ol

5880000 5890000 5900000 5910000 5920000 5930000

010

2030

4050

6070

CJuvalLaatsch

FinailTerz

Vioz/Coro

Figure 6: Example of a candidate gene (AT5G17860, encoding for CAX7) in which allele fre-quences were highly differentiated between high and low altitude populations. (A) Plot of thegenome-wide relative rank of the Z statistics. High values indicate a strong differentiation. (B)Plot of the relative allele frequencies of the two alleles segregating at a polymorphic site. Theheight of the bar is proportional to the allele frequency. (C) Sequence coverage of the genomicregion in the five different populations.

17


https://doi.org/10.1101/023051

Discussion423

Presence of phenotypic variation in low and high altitude populations We identified424

phenotypic variation in two ecologically relevant traits in the five populations but their spatial425

distribution did not show a strong correlation with altitude. Due to the large difference in426

altitude between populations (>1,000 m) and a decrease of average temperature by 0.6 °C427

per 100 m altitude (Barry, 1992) we expected strong selection for freezing tolerance at high428

altitude. Although we noted a significant variation in freezing tolerance, populations with429

similar frost resistance levels did not originate from the same altitude or water drainage (river)430

system. A possible explanation is that selection for cold tolerance is mainly influenced by431

microclimatic conditions which may be strongly influenced by local topology. The high altitude432

population Vioz/Coro is found on South-exposed rocky outcrops at protected sites underneath433

rock overhangs. Temperature logger data from the two high altitude sites suggest that the434

Vioz/Coro site is less prone to freezing during winter (Figure SS2). In addition, high alpine A.435

thaliana populations may switch to a summer annual life habit allowing the plants to avoid the436

period with high risk of freezing damage (Luo et al , 2014). Our observations of plantlets at the437

different localities suggest that the Vioz/Coro population consists of summer annuals, while at438

the other four sites plantlets germinate in the fall and persist throughout the winter.439

The spatial pattern of response to light and UV-B stress differed from frost tolerance. The440

population at the highest altitude (Vioz/Coro) showed a lack of anthocyanin accumulation441

under light stress. Anthocyanins protect against cell damage by light-induced photooxidation442

(Chalker-Scott, 1999; Powles, 1984) and the plants from the Vioz/Coro population exhibited443

tissue bleaching in old roset leafs under light stress and were more sensitive to extended elevated444

UV-B levels. Flavonoid-deficient mutants of A. thaliana show a similar phenotype (Li et al ,445

1993). Solar radiation at open sky and fluctuations in radiation intensity increase with higher446

altitude (Blumthaler et al , 1997; Korner, 2007) and high altitude populations need to adapt to447

more intense UV-B and light conditions. For example, the Shahkdara accession of A. thaliana448

was collected at 3,063 m a.s.l. and exhibited a constitutive UV-B protection (Biswas & Jansen,449

2012). We therefore suggest that the deficiency in anthocyanin accumulation in the Vioz/Coro450

population is not an adaptation to high altitudes but a loss-of-function phenotype resulting451

from genetic drift at the range margins (Eckert et al , 2008).452

In summary, the spatial patterns of variation in the two traits do not support a simple453

pattern of adaptation to altitude because we expected both high altitude populations to show454

a similar phenotype and the demographic analysis suggested that there was sufficient time455

available for adaptation to high altitude. Possible explanations are that a difference of 1,700456

m altitude between sampling locations does not represent a strong selection gradient or that457

the microenvironment or random drift and metapopulation dynamics of small demes confounds458

clinal trait variation.459

Patterns of genetic diversity and population differentiation In A. thaliana, levels460

of genetic diversity differ strongly among populations at small and large geographical scales461

18


https://doi.org/10.1101/023051

(Bomblies et al , 2010; Cao et al , 2011). A low genetic diversity limits the potential to respond462

to selection and may lead to the accumulation of deleterious alleles (Alleaume-Benharira et al ,463

2006). In our sample, both high altitude populations exhibited reduced genetic diversity and464

genome-wide negative Tajima’s D values that indicate a deviation from a standard neutral465

model. This pattern may result from bottlenecks with subsequent population growth (Simonsen466

et al , 1995) as expected after the colonization of new habitats like high altitude sites at the outer467

vertical limit of a species distribution. However, we did not observe an increase of deleterious468

amino acid polymorphisms in high altitude populations that would accompany such a bottleneck469

due to a reduction of purifying selection (Gunther & Schmid, 2010).470

Although diversity estimates differed significantly among the five populations, we found471

little evidence for a strong population structure. The average pairwise FST value of 0.128 was472

three times smaller than in local populations near Tubingen, Germany (FST = 0.52; Bomblies473

et al , 2010) and smaller than the smallest observed pairwise FST in a sample of Swiss alpine A.474

thaliana populations that were based on 24 SSR loci (Luo et al , 2014). In addition, pairwise475

FST values were very similar for all pairs of populations, consistent with a low number of unique476

SNPs per accession and a high level of LD in a different sample of South Tyrolian accessions477

(Cao et al , 2011). The demographic history of numerous alpine plant species was influenced478

by glaciation which caused repeated cycles of retreats into refugia and colonization of newly479

available habitats in interglacial periods (Schonswetter et al , 2005). Refugial populations in480

alpine valleys or local refugia on mountain tops (nunataks) above glaciers may have contributed481

to a strong genetic differentiation between subpopulations. A low differentiation and a star-like482

relationship with gene flow between populations (e.g., the low-altitude Juval and the high-483

altitude Finail populations; Figure 5) of our five populations is not consistent with such a484

scenario and instead suggest that these sites were colonized recently from refugia at the edge or485

outside the alps. On the other hand, the two high altitude populations were more closely related486

to each other than to the respectice low-altitude populations from the same river drainage. This487

may reflect an origin from the same (mountaintop) refugium or a recent dispersal by animals.488

Generally, a rapid postglacial recolonization of alpine habitats by A. thaliana is possible because489

of the mainly self-fertilizing mode of reproduction and dispersal by animals like mountain goats.490

The high altitude populations were found only at highly disturbed resting sites of mountain491

goats.492

To investigate the age of the split of low and high-altitude populations, we estimated demo-493

graphic parameters for different demographic models with ABC (Beaumont et al , 2002). Both494

pairs of low and high altitude population supported a model with a smaller effective population495

size in high altitude populations caused by a bottleneck and subsequent population growth. The496

90% HDPI intervals for population split times ranged between 14,000 and 260,000 years before497

present. This includes time before and after the last glacial maximum (LGM) as deglaciation498

on the Northern Hemisphere began about 19-20 thousand years ago (Clark et al , 2009). Con-499

sequently, our ABC analysis did not allow to answer the question whether the split between500

low- and high-altitude populations occurred before or after the LGM, though the probability is501

19


https://doi.org/10.1101/023051

higher that it occurred before. The wide HDPI 90 interval may result from a limitation of pool502

sequencing because summary statistics used for ABC were limited to measures of gene frequen-503

cies and did not include haplotype-based statistics, which provide additional power for inferring504

population split times and admixture levels by quantifying the length of shared fragments (Sved505

et al , 2008; Sankararaman et al , 2012; Hellenthal et al , 2014).506

Patterns of genetic variation and local selective sweeps In A.thaliana only few species-507

wide selective sweeps were found so far (Clark et al , 2007; Childs et al , 2010; Cao et al , 2011;508

Horton et al , 2012), and the overall frequency of adaptive evolution is very low or absent based on509

genome-wide polymorphism data (Gossmann et al , 2010; Slotte et al , 2011) although evidence510

for local adaptation was found in several studies (Fournier-Level et al , 2011; Hancock et al ,511

2011; Agren & Schemske, 2012; Mendez-Vigo et al , 2011; Long et al , 2013; Huber et al , 2014).512

Therefore, searching for signatures of selection at smaller geographic scales along environmental513

gradients may identify phenotypic and genomic footprints of adaptive evolution. We identified514

candidate genes for high-altitude adaptation with an outlier approach. It should be noted that515

this is no formal test of selection and candidate genes require further validation (Pavlidis et al ,516

2012).517

Patterns of genetic variation at some candidate genes suggested that targets of selection were518

difficult to elucidate because large genomic regions (>50k) around significantly associated SNPs519

involving multiple adjacent genes showed a population-specific sweep pattern (Figures 6, S7 and520

S8). The comparison of the three genomic regions showed that despite an overall low level of521

genetic diversity in the alpine populations, complex patterns of genetic diversity that involved522

presence/absence variants (PAVs) and regions of high nucleotide diversity were observed. For523

example, in our sample, a single SNP was significantly differentiated between low and high524

altitude populations and some neighboring SNPs also showed a clinal pattern (Figures 6). It was525

previously identified in Arabidopsis lyrata as a candidate selection target for adaptation to soils526

with a low Ca:Mg-ratio (Turner et al , 2008) and A. thaliana is generally found on siliceous rock527

with low Ca content. However, in the direct neighbourhood, although pool sequencing did not528

allow the analysis of haplotypes, patterns of sequence coverage strongly suggested that a PAV529

comprising two well characterized NBS-LRR genes, CSA1 and CHS3, may have contributed530

to local adaptation since the deletion was fixed in the high altitude and Laatsch populations,531

whereas it segregated in the low-altitude populations (Figure 6). In addition both genes showed532

a PAV across the whole species because CSA1 was not covered by sequence reads in 11 and533

CHS3 in 9 out of 18 high coverage reference genomes of a diverse set of accessions (Gan et al ,534

2011). PAVs are abundant across the genome of A.thaliana (Tan et al , 2012), particularly in535

resistance genes that include NBS-LRR genes (Bergelson et al , 2001; Shen et al , 2006; Guo et al ,536

2011; Karasov et al , 2014). It remains open whether selection or genetic drift contributed to the537

fixation at these genes. They are involved in pathogen response, but CSA1 additionally plays a538

role in shade avoidance and response to red light (Faigon-Soverna et al , 2006) and knockouts of539

CHS3 cause chilling-sensitive plants (Schneider et al , 1995). A comparison of sequence coverage540

pattern against all genes annotated in TAIR10 showed, however, that a similar deletion pattern541

20


https://doi.org/10.1101/023051

was observed in only 15 genes that included other NBS-LRR genes or genes of unknown function.542

The pattern of diversity at the AtEHD1 gene (AT3G20290) may represent a population-543

specific selective sweep because of numerous SNPs with a high Z-score in this region. The544

same alleles are fixed in the two high-altitude populations and the diversity is strongly reduced545

over a range of about 50 kb (Figure S7). In contrast, the low altitude populations appeared to546

be highly polymorphic in this region. The identification of putative selection targets requires547

further study, but it is noteworthy that AtEHD1 increases tolerance to salt stress (Bar et al ,548

2013).549

We observed another pattern of genetic diversity in the gene encoding the purple acid phos-550

phatase 15 (AT3G07130) where only the highest population was differentiated (Figure S8).551

This gene may affect phosphorus efficiency by remobilizing inorganic phosphate from organic552

phosphate sources (Wang et al , 2009) and is a plausible selection target because phosphorus553

was highest in the highest population (Table S5). Taken together, diversity patterns in several554

genomic regions suggest that adaptation to soil conditions may have played a role in the history555

of these populations (Baxter et al , 2010).556

Changes in light conditions and temperature differences are among the expected environ-557

mental differences between altitudes. We did not identify genes known to be involved in frost558

resistance in the GO enrichment analysis (Table S7), which corresponds to the patterns of pheno-559

typic differentiation. However, genes associated with heat acclimation were among the identified560

candidates, which may relate to the observation that high temperatures in late spring terminate561

the reproductive cycle in all low altitude populations by causing plant death. In contrast, mild562

temperatures at high altitudes allow reproduction throughout the summer suggesting selection563

for heat acclimation may be relaxed (personal observation: C. Lampei). Furthermore, in several564

genes involved in shade avoidance and response to light stimulus as well as mutations in the565

anthocyanin (Table S8), flavonol and phenylpropanoid biosynthesis pathways (Table S7) were566

correlated with altitude. This pattern is consistent with phenotypic differentiation, although567

the anthocyanin deficiency may constitute a loss of function mutation in the highest population,568

indicating that genetic differentiation at some genes may not result from past selection.569

In a recent genome-wide scan in A. halleri, genes associated with red or far red light response570

were differentiated with altitude in two independent altitude gradients in Japan (Kubota et al ,571

2015). Further, genes involved in defence response were also enriched in an Alpine altitude572

gradient study in the same plant (Fischer et al , 2013). Notably, the enrichment analysis did573

not identify any soil related categories among the nominally significant annotations, which is574

in strong contrast to the annotation of the top 25 highly differentiated genes. This corresponds575

to the observation by Fischer et al (2013) that the GO enrichment analysis did not match with576

outlier correlations with environmental variables.577

Methods to identify selection along altitudinal gradients Pool sequencing is a cost-578

efficient method to characterize genetic variation using allele frequency estimates. Reliable579

estimates of population-specific allele frequencies require a sufficient sample size and sequence580

coverage (Schlotterer et al , 2014). Our sequenced pools fulfill recommendations for pool sequenc-581

21


https://doi.org/10.1101/023051

ing and we expect that the observed patterns of genetic diversity resemble genomic differences582

and not artefacts of the method. Both our phenotypic and genomic analysis identified substan-583

tial variation along altitudinal gradients. Studies of local adaptation need to consider the effect584

of sampling design on the probability to identify false positives (Meirmans, 2015). Simulations585

suggest that studies of local adaptation should be based on several pairs of populations that586

differ in the environmental variable of interest (Lotterhos & Whitlock, 2015). We included two587

population pairs to have to samples of putative adaptive processes and demographic history.588

The closer relationship of the high-altitude populations, however, suggest that genetic variation589

shares a joint history and are not entirely independent. For this reason, the power to identify590

targets of altitudinal adaptation increases with more population pairs, if local populations are591

small and isolated and therefore prone to genetic drift. Future studies of altitudinal adaptation592

in A. thaliana should therefore be based on a denser sampling of populations and include ad-593

ditional phenotypic traits (Montesinos-Navarro et al , 2011; Luo et al , 2014). Given the patchy594

and highly disjunct distribution of A. thaliana in the high Alps, this might require system-595

atic sampling over a larger geographic area to disentangle demography from natural selection596

(Meirmans et al , 2011).597

Conclusion The phenotypic variation in response to two environmental variables, which pre-598

dictably change with altitude, was not consistent with selection for high-altitude adaptation in599

our A. thaliana populations from the North Italian Alps. Microevolutionary responses to local600

topology at the population sites, founder effects and small population sizes at high altitude601

may overrule mean altitudinal change in environmental variables. In contrast, several genomic602

regions showed patterns of genetic variation consistent with population-specific selection-driven603

sweeps, that were associated with observed environmental differences (e.g. soil parameters).604

Furthermore, the detection of outlier genes suggested some plausible targets positive selection.605

These candidate genes can be tested at a functional level in future studies to test their potential606

role in adaptation to high altitude. Patterns of genetic diversity indicated a closer relation-607

ship of the high altitude populations that may reflect a common demographic history during608

glacial-interglacial cycles.609

Acknowledgements610

We thank Elisabeth Kokai-Kota for lab assistance, Hinrich Bremer for access to a conductivity611

meter, Dr. Nikolaus Merkt for access to a UV-light meter and Fabian Freund for statistical dis-612

cussions. This work was funded by the ESF EUROCORES project EpiCol (SCHM 1354/4-1),613

the German Federal Ministry for Education and Research (BMBF) within the AgroClustEr Syn-614

breed—Synergistic plant and animal breeding (FKZ0315528D), and BMBF Plant2030 project615

RYE-SELECT (FKZ 0315946E).616

22


https://doi.org/10.1101/023051

References617

Agren J, Schemske DW (2012) Reciprocal transplants demonstrate strong adaptive differenti-618

ation of the model organism Arabidopsis thaliana in its native range. The New Phytologist,619

194, 1112–22.620

Al-Shehbaz Ia, O’Kane SL (2002) Taxonomy and phylogeny of Arabidopsis (Brassicaceae). The621

Arabidopsis book / American Society of Plant Biologists, 1, e0001.622

Alleaume-Benharira M, Pen IR, Ronce O (2006) Geographical patterns of adaptation within a623

species’ range: interactions between drift and gene flow. Journal of Evolutionary Biology, 19,624

203–15.625

Altschul SF, Madden TL, Schaffer AA, et al (1997) Gapped BLAST and PSI-BLAST: a new626

generation of protein database search programs. Nucleic Acids Research, 25, 3389–3402.627

Ashburner M, Ball CA, Blake JA, et al (2000) Gene ontology: tool for the unification of biology.628

The Gene Ontology Consortium. Nature Genetics, 25, 25–29.629

Bar M, Leibman M, Schuster S, Pitzhadza H, Avni A (2013) EHD1 functions in endosomal630

recycling and confers salt tolerance. PloS ONE, 8, e54533.631

Barry RG (1992) Mountain Weather and Climate. Routledge, London, 2nd edn..632

Baxter I, Brazelton JN, Yu D, et al (2010) A coastal cline in sodium accumulation in Arabidopsis633

thaliana is driven by natural variation of the sodium transporter AtHKT1; 1. PLoS Genetics,634

6, e1001193.635

Beaumont Ma, Zhang W, Balding DJ (2002) Approximate Bayesian computation in population636

genetics. Genetics, 162, 2025–35.637

Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful638

approach to multiple testing. Journal of the Royal Statistical Society. Series B, 57, 289–300.639

Bergelson J, Kreitman M, Stahl EA, Tian D (2001) Evolutionary dynamics of plant R-genes.640

Science, 292, 2281–5.641

Bergelson J, Roux F (2010) Towards identifying genes underlying ecologically relevant traits in642

Arabidopsis thaliana. Nature Reviews Genetics, 11, 867–879.643

Biswas DK, Jansen MAK (2012) Natural variation in UV-B protection amongst Arabidopsis644

thaliana accessions. Emirates Journal of Food and Agriculture, 24, 621–631.645

Blumthaler M, Ambach W, Ellinger R (1997) Increase in solar UV radiation with altitude.646

Journal of Photochemistry and Photobiology B: Biology, 39, 130–134.647

23


https://doi.org/10.1101/023051

Boitard S, Schlotterer C, Nolte V, Pandey RV, Futschik A (2012) Detecting selective sweeps648

from pooled next-generation sequencing samples. Molecular Biology and Evolution, 29, 2177–649

86.650

Bomblies K, Yant L, Laitinen Ra, et al (2010) Local-scale patterns of genetic variability, out-651

crossing, and spatial structure in natural stands of Arabidopsis thaliana. PLoS Genetics, 6,652

e1000890.653

Cao J, Schneeberger K, Ossowski S, et al (2011) Whole-genome sequencing of multiple Ara-654

bidopsis thaliana populations. Nature Genetics, 43, 956–963.655

Chalker-Scott L (1999) Environmental significance of anthocyanins in plant stress responses.656

Photochemistry and Photobiology, 70, 1–9.657

Childs LH, Witucka-Wall H, Gunther T, et al (2010) Single feature polymorphism (SFP)-based658

selective sweep identification and association mapping of growth-related metabolic traits in659

Arabidopsis thaliana. BMC Genomics, 11, 188.660

Cingolani P, Platts A, Wang LL, et al (2012) A program for annotating and predicting the effects661

of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster662

strain w1118; iso-2; iso-3. Fly, 6, 80–92.663

Clark PU, Dyke AS, Shakun JD, et al (2009) The Last Glacial Maximum. Science, 325, 710–664

714.665

Clark RM, Schweikert G, Toomajian C, et al (2007) Common sequence polymorphisms shaping666

genetic diversity in Arabidopsis thaliana. Science, 317, 338–342.667

Cooper L, Walls RL, Elser J, et al (2013) The plant ontology as a tool for comparative plant668

anatomy and genomic analyses. Plant & Cell Physiology, 54, e1.669

Cox MP, Peterson Da, Biggs PJ (2010) SolexaQA: At-a-glance quality assessment of Illumina670

second-generation sequencing data. BMC Bioinformatics, 11, 485.671

Csillery K, Francois O, Blum MGB (2012) abc: an R package for approximate Bayesian com-672

putation (ABC). Methods in Ecology and Evolution, 3, 475–479.673

DevelopmentCoreTeam (2014) R: A Language and Environment for Statistical Computing. Vi-674

enna.675

Eckert C, Samis K, Lougheed S (2008) Genetic variation across species’ geographical ranges:676

the central–marginal hypothesis and beyond. Molecular Ecology, 17, 1170–1188.677

Etterson JR, Shaw RG (2001) Constraint to adaptive evolution in response to global warming.678

Science, 294, 151–154.679

24


https://doi.org/10.1101/023051

Fabian DK, Kapun M, Nolte V, et al (2012) Genome-wide patterns of latitudinal differentiation680

among populations of Drosophila melanogaster from North America. Molecular Ecology, 21,681

4748–69.682

Faigon-Soverna A, Harmon FG, Storani L, et al (2006) A constitutive shade-avoidance mutant683

implicates TIR-NBS-LRR proteins in Arabidopsis photomorphogenic development. The Plant684

Cell, 18, 2919–28.685

Fischer MC, Rellstab C, Tedder A, et al (2013) Population genomic footprints of selection686

and associations with climate in natural populations of Arabidopsis halleri from the Alps.687

Molecular Ecology, pp. 5594–5607.688

Fournier-Level a, Korte a, Cooper MD, Nordborg M, Schmitt J, Wilczek AM (2011) A map of689

local adaptation in Arabidopsis thaliana. Science, 334, 86–9.690

Friedman N, Ninio M, Pe’er I, Pupko T (2002) A structural EM algorithm for phylogenetic691

inference. Journal of Computational Biology, 9, 331–353.692

Futschik A, Schlotterer C (2010) The next generation of molecular markers from massively693

parallel sequencing of pooled DNA samples. Genetics, 186, 207–18.694

Gan X, Stegle O, Behr J, et al (2011) Multiple reference genomes and transcriptomes for695

Arabidopsis thaliana. Nature, 477, 419–23.696

Gossmann TI, Song BH, Windsor AJ, et al (2010) Genome wide analyses reveal little evidence697

for adaptive evolution in many plant species. Molecular Biology and Evolution, 27, 1822–32.698

Gunther T, Coop G (2013) Robust Identification of Local Adaptation from Allele Frequencies.699

Genetics, 195, 205–220.700

Gunther T, Schmid KJ (2010) Deleterious amino acid polymorphisms in Arabidopsis thaliana701

and rice. Theoretical and Applied Genetics, 121, 157–68.702

Guo YL, Fitz J, Schneeberger K, Ossowski S, Cao J, Weigel D (2011) Genome-wide comparison703

of nucleotide-binding site-leucine-rich repeat-encoding genes in Arabidopsis. Plant Physiology,704

157, 757–69.705

Hall B (2013) LaplacesDemon: An R Package for Bayesian Inference, version 13.11.17.706

Hancock AM, Brachi B, Faure N, et al (2011) Adaptation to climate across the Arabidopsis707

thaliana genome. Science, 334, 83–6.708

He Z, Zhai W, Wen H, et al (2011) Two evolutionary histories in the genome of rice: the roles709

of domestication genes. PLoS Genetics, 7, e1002100.710

Hellenthal G, Busby GBJ, Band G, et al (2014) A genetic atlas of human admixture history.711

Science, 343, 747–751.712

25


https://doi.org/10.1101/023051

Horton MW, Hancock AM, Huang YS, et al (2012) Genome-wide patterns of genetic variation713

in worldwide Arabidopsis thaliana accessions from the RegMap panel. Nature Genetics, 44,714

212–6.715

Hu TT, Pattyn P, Bakker EG, et al (2011) The Arabidopsis lyrata genome sequence and the716

basis of rapid genome size change. Nature Genetics, 43, 476–481.717

Huber CD, Nordborg M, Hermisson J, Hellmann I (2014) Keeping it local: Evidence for positive718

selection in Swedish Arabidopsis thaliana. Molecular Biology and Evolution, 31, 3026–3039.719

Huxley JS (1938) Clines: an auxiliary method in taxonomy. Nature, 142, 219–220.720

Kanehisa M, Goto S, Sato Y, Furumichi M, Tanabe M (2012) KEGG for integration and inter-721

pretation of large-scale molecular data sets. Nucleic Acids Research, 40, gkr988.722

Karasov TL, Horton MW, Bergelson J (2014) Genomic variability as a driver of plant-pathogen723

coevolution? Current Opinion in Plant Biology, 18, 24–30.724

Kim S, Plagnol V, Hu TT, et al (2007) Recombination and linkage disequilibrium in Arabidopsis725

thaliana. Nature Genetics, 39, 1151–1155.726

Kofler R, Betancourt AJ, Schlotterer C (2012) Sequencing of pooled DNA samples (Pool-Seq)727

uncovers complex dynamics of transposable element insertions in Drosophila melanogaster.728

PLoS Genetics, 8, –1002487.729

Kofler R, Orozco-terWengel P, De Maio N, et al (2011a) PoPoolation: A toolbox for population730

genetic analysis of next generation sequencing data from pooled individuals. PLoS ONE, 6,731

e15925.732

Kofler R, Pandey RV, Schlotterer C (2011b) PoPoolation2: identifying differentiation between733

populations using sequencing of pooled DNA samples (Pool-Seq). Bioinformatics, 27, 3435–6.734

Kolaczkowski B, Kern AD, Holloway AK, Begun DJ (2011) Genomic differentiation between735

temperate and tropical Australian populations of Drosophila melanogaster. Genetics, 187,736

245–60.737

Koornneef M, Alonso-Blanco C, Vreugdenhil D (2004) Naturally occurring genetic variation in738

Arabidopsis thaliana. Annual Review of Plant Biology, 55, 141–72.739

Korner C (2007) The use of ’altitude’ in ecological research. Trends in Ecology & Evolution,740

22, 569–74.741

Kubota S, Iwasaki T, Hanada K, et al (2015) A Genome Scan for Genes Underlying742

Microgeographic-Scale Local Adaptation in a Wild Arabidopsis Species. PLOS Genetics,743

11, e1005361.744

Lai L, Liberzon A, Hennessey J, et al (2012) AraPath: a knowledgebase for pathway analysis745

in Arabidopsis. Bioinformatics, 28, 2291–2.746

26


https://doi.org/10.1101/023051

Lamesch P, Berardini TZ, Li D, et al (2012) The Arabidopsis Information Resource (TAIR):747

improved gene annotation and new tools. Nucleic Acids Research, 40, D1202–D1210.748

Lamichhaney S, Martinez Barrio A, Rafati N, et al (2012) Population-scale sequencing reveals749

genetic differentiation due to local adaptation in Atlantic herring. Proceedings of the National750

Academy of Sciences of the United States of America, 109, 19345–50.751

Li H, Durbin R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform.752

Bioinformatics, 25, 1754–60.753

Li H, Handsaker B, Wysoker A, et al (2009) The Sequence Alignment/Map format and SAM-754

tools. Bioinformatics, 25, 2078–9.755

Li J, Ou-Lee TM, Raba R, Amundson RG, Last RL (1993) Arabidopsis Flavonoid Mutants Are756

Hypersensitive to UV-B Irradiation. The Plant Cell, 5, 171–179.757

Long Q, Rabanal Fa, Meng D, et al (2013) Massive genomic variation and strong selection in758

Arabidopsis thaliana lines from Sweden. Nature Genetics, 45, 884–90.759

Lotterhos K, Whitlock MC (2015) The relative power of genome scans to detect local adaptation760

depends on sampling design and statistical method. Molecular Ecology, 24, 1031–1046.761

Luo Y, Dong X, Yu TY, et al (2015) A single nucleotide deletion in GA20ox1 causes alpine762

dwarfism in Arabidopsis thaliana. Plant Physiology, pp. pp–00005.763

Luo Y, Widmer A, Karrenberg S (2014) The roles of genetic drift and natural selection in764

quantitative trait divergence along an altitudinal gradient in Arabidopsis thaliana. Heredity,765

pp. 220–228.766

Manel S, Gugerli F, Thuiller W, et al (2012) Broad-scale adaptive genetic variation in alpine767

plants is driven by temperature and precipitation. Molecular Ecology, 21, 3729–38.768

Martin M, Gavazov K, Korner C, Hattenschwiler S, Rixen C (2010) Reduced early growing769

season freezing resistance in alpine treeline plants under elevated atmospheric CO 2. Global770

Change Biology, 16, 1057–1070.771

Mayr E (1942) Systematics and the Origin of Species, from the Viewpoint of a Zoologist. Harvard772

University Press, New York.773

Meirmans PG (2015) Seven common mistakes in population genetics and how to avoid them.774

Molecular Ecology.775

Meirmans PG, Goudet J, Gaggiotti OE (2011) Ecology and life history affect different aspects776

of the population structure of 27 high-alpine plants. Molecular Ecology, 20, 3144–3155.777

Mendez-Vigo B, Pico FX, Ramiro M, Martınez-Zapater JM, Alonso-Blanco C (2011) Altitudinal778

and climatic adaptation is mediated by flowering traits and FRI, FLC, and PHYC genes in779

Arabidopsis. Plant Physiology, 157, 1942–55.780

27


https://doi.org/10.1101/023051

Montesinos-Navarro A, Wig J, Xavier Pico F, Tonsor SJ (2011) Arabidopsis thaliana populations781

show clinal variation in a climatic gradient associated with altitude. The New Phytologist,782

189, 282–94.783

Mueller LA, Zhang P, Rhee SY (2003) AraCyc: a biochemical pathway database for Arabidopsis.784

Plant Physiology, 132, 453–60.785

Orozco-terWengel P, Kapun M, Nolte V, Kofler R, Flatt T, Schlotterer C (2012) Adaptation of786

Drosophila to a novel laboratory environment reveals temporally heterogeneous trajectories787

of selected alleles. Molecular Ecology, 21, 4931–41.788

Ossowski S, Schneeberger K, Lucas-Lledo JI, et al (2010) The rate and molecular spectrum of789

spontaneous mutations in Arabidopsis thaliana. Science, 327, 92–94.790

Paradis E, Claude J, Strimmer K (2004) APE: Analyses of phylogenetics and evolution in R791

language. Bioinformatics, 20, 289–90.792

Park JS, Choung MG, Kim JB, et al (2007) Genes up-regulated during red coloration in UV-B793

irradiated lettuce leaves. Plant Cell Reports, 26, 507–16.794

Patterson N, Moorjani P, Luo Y, et al (2012) Ancient admixture in human history. Genetics,795

192, 1065–1093.796

Pavlidis P, Jensen JD, Stephan W, Stamatakis A (2012) A critical assessment of storytelling:797

gene ontology categories and the importance of validating genomic scans. Molecular Biology798

and Evolution, 29, 3237–48.799

Pavlidis P, Laurent S, Stephan W (2010) msABC: a modification of Hudson’s ms to facilitate800

multi-locus ABC analysis. Molecular Ecology Resources, 10, 723–727.801

Pickrell JK, Pritchard JK (2012) Inference of population splits and mixtures from genome-wide802

allele frequency data. PLoS Genetics, 8, e1002967.803

Pinheiro J, Bates D, DebRoy S, Sarkar D, Team RC (2014) nlme: Linear and Nonlinear Mixed804

Effects Models. R package version 3.1-117.805

Powles SB (1984) Photoinhibition of Photosynthesis Induced by Visible Light. Annual Review806

of Plant Physiology, 35, 15–44.807

Reich D, Thangaraj K, Patterson N, Price AL, Singh L (2009) Reconstructing Indian population808

history. Nature, 461, 489–94.809

Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylo-810

genetic trees. Molecular Biology and Evolution, 4, 406–25.811

Sankararaman S, Patterson N, Li H, Paabo S, Reich D (2012) The date of interbreeding between812

Neandertals and modern humans. PLoS Genetics, 8, e1002947.813

28


https://doi.org/10.1101/023051

Schindelin J, Arganda-Carreras I, Frise E, et al (2012) Fiji: an open-source platform for814

biological-image analysis. Nature methods, 9, 676–682.815

Schlotterer C, Tobler R, Kofler R, Nolte V (2014) Sequencing pools of individuals — mining816

genome-wide polymorphism data without big funding. Nature Reviews Genetics, 15, 749–763.817

Schneider JC, Suzanne H, Somerville CR (1995) Chilling-sensitive mutants of Arabidopsis. Plant818

Molecular Biology Reporter, 13, 11–17.819

Schonswetter P, Stehlik I, Holderegger R, Tribsch A (2005) Molecular evidence for glacial refugia820

of mountain plants in the European Alps. Molecular Ecology, 14, 3547–55.821

Shen J, Araki H, Chen L, Chen JQ, Tian D (2006) Unique evolutionary mechanism in R-genes822

under the presence/absence polymorphism in Arabidopsis thaliana. Genetics, 172, 1243–50.823

Simonsen KL, Churchill GA, Aquadro CF (1995) Properties of statistical tests of neutrality for824

DNA polymorphism data. Genetics, 141, 413–29.825

Slotte T, Bataillon T, Hansen TT, St Onge K, Wright SI, Schierup MH (2011) Genomic determi-826

nants of protein evolution and polymorphism in Arabidopsis. Genome Biology and Evolution,827

3, 1210–9.828

Stinchcombe JR, Weinig C, Ungerer M, et al (2004) A latitudinal cline in flowering time in829

Arabidopsis thaliana modulated by the flowering time gene FRIGIDA. Proceedings of the830

National Academy of Sciences of the United States of America, 101, 4712–7.831

Stone EA, Sidow A (2005) Physicochemical constraint violation by missense substitutions me-832

diates impairment of protein function and disease severity. Genome Research, 15, 978–986.833

Sved JA, McRae AF, Visscher PM (2008) Divergence between human populations estimated834

from linkage disequilibrium. American Journal of Human Genetics, 83, 737–743.835

Tan S, Zhong Y, Hou H, Yang S, Tian D (2012) Variation of presence/absence genes among836

Arabidopsis populations. BMC Evolutionary Biology, 12, 86.837

Thiel-Egenter C, Alvarez N, Holderegger R, et al (2011) Break zones in the distributions of838

alleles and species in alpine plants. Journal of Biogeography, 38, 772–782.839

Turner TL, Bourne EC, Von Wettberg EJ, Hu TT, Nuzhdin SV (2010) Population resequencing840

reveals local adaptation of Arabidopsis lyrata to serpentine soils. Nature Genetics, 42, 260–3.841

Turner TL, von Wettberg EJ, Nuzhdin SV (2008) Genomic analysis of differentiation between842

soil types reveals candidate genes for local adaptation in Arabidopsis lyrata. PloS ONE, 3,843

e3183.844

Wang X, Wang Y, Tian J, Lim BL, Yan X, Liao H (2009) Overexpressing AtPAP15 enhances845

phosphorus efficiency in Soybean. Plant Physiology, 151, 233–40.846

29


https://doi.org/10.1101/023051

Wu CH, Apweiler R, Bairoch A, et al (2006) The Universal Protein Resource (UniProt): an847

expanding universe of protein information. Nucleic Acids Research, 34, D187–D191.848

Yilmaz A, Mejia-Guerra MK, Kurz K, Liang X, Welch L, Grotewold E (2011) AGRIS: the849

Arabidopsis Gene Regulatory Information Server, an update. Nucleic Acids Research, 39,850

D1118–D1122.851

Zhu Y, Bergland AO, Gonzalez J, Petrov DA (2012) Empirical validation of pooled whole852

genome population re-sequencing in Drosophila melanogaster. PLoS ONE, 7, e41901.853

30


https://doi.org/10.1101/023051

Data availability The raw sequence data were deposited in the Short Read Archive (XXXXX).854

SNP calls and other downstream sequence analyse data were deposited with with all phenotypic855

measurements and the R and Python scripts used for the analysis at DataDryad under accession856

number XXXXX.857

31


https://doi.org/10.1101/023051

Phenotypic and genomic differentiation along altitudinal1

gradients in South-Alpine populations of Arabidopsis thaliana2

Supporting information3

Torsten Gunther$,‡,§, Christian Lampei$,‡, Ivan Barilar$, and Karl J. Schmid$,∗4

July 22, 20155

$ Institute of Plant Breeding, Seed Science and Population Genetics, University of Hohenheim,6

Stuttgart, Germany7

§Department of Evolutionary Biology, EBC, Uppsala University, Uppsala, Sweden.8

∗Corresponding author: E-mail: [email protected]

‡These authors contributed equally to this work.10

Supplementary Methods11

Resistance to high UV-B radiation To test resistance to higher UV-B dosages we sub-12

jected 44 days old plants (10 pots and genotypes per population) which were still small due13

to 19 days of vernalization (4� and short day (8h light)) to increasing UV-B dosages of con-14

tinuous UV-B light over 3 days. The UV-B levels were 0, 2.47, 6.39 and 15.16 kJm−2day−1.15

After this treatment plants were left to recover for 8 days to improve the separation of dead16

and alive biomass. All plants were evaluated manually with their identity hidden from the17

investigator. Dead and living tissue was separated for each pot and oven dried for 12 h at 9018

� before weighting. A GLM with dead biomass odds as response variable and the fixed effects19

population and UV-B dosage was fitted with a quasibinomial error distribution (logit link) in20

R (DevelopmentCoreTeam, 2014, package:stats).21

Soil properties In all sites soil samples of the top soil were collected. The soil samples were22

composed from about 40 individual samples per population site which were collected with a23

hand shovel from the uppermost 5 cm well spread across the distribution of the population.24

The soil was mixed, homogenized and analysed for standard parameters (Table S5).25

1


https://doi.org/10.1101/023051

Supplementary Figures26

UV−B dosage (kJ m−2 day−1)

Pro

port

ion

dead

bio

mas

s pe

r pl

ant

0.0

0.2

0.4

0.6

0.8

1.0

600 Juval 850 Terz1000 Laatsch2260 Finail2355 Vioz/Coro

Altitude, Population

0 2.5 6.4 15.2

ab

abb

a

b

Figure S1: Proportion of dead tissue after 3 days continuous UV-B treatment with differentdosages. Letters above error bars indicate homogeneous subsets after correction for false dis-covery rate for the highest UV-B dosage

2


https://doi.org/10.1101/023051

010

020

030

0

01020304050

Day

in y

ear

Temperature (° C)

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

tO

ktN

ovD

ecJa

n

Vio

z/C

oro

Fin

ail

Fig

ure

S2:

Dai

lym

inim

um

and

max

imu

mte

mp

erat

ure

for

the

top

soil

laye

r(<

5cm

)in

the

two

hig

halt

itu

de

site

s

3


https://doi.org/10.1101/023051

Figure S3: Four demographic models without and with migration investigated in this study.NOGROWTH - Model without the growth of the high altitude population after the split.STEPGROWTH - Model with one instantaneous size change of the high altitude populationafter the split. GROWTH - Model with simulated growth of the high altitude population afterthe split. EXPONENTIAL - Model with an exponential growth of the high altitude populationsometime after the split. Ts - split time. Nlow and Nhigh - effective population size of the lowand high altitude population, respectively.

4


https://doi.org/10.1101/023051

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95

Juval

Derived allele frequency

0.0

0.2

0.4

0.6

NON_SYNONYMOUS_CODINGSTART_LOSTSTOP_GAINEDSTOP_LOSTSYNONYMOUS_CODING

INTRONUTR_5_PRIMEUTR_3_PRIMEINTERGENIC

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95

Laatsch


0.0

0.2

0.4

0.6



0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95

Finail


0.0

0.2

0.4

0.6



0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95

Terz


0.0

0.1

0.2

0.3

0.4

0.5

0.6 NON_SYNONYMOUS_CODING

START_LOSTSTOP_GAINEDSTOP_LOSTSYNONYMOUS_CODING


0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95

Vioz/Coro


0.0

0.2

0.4

0.6



Figure S4: Derived allele frequency distributions across all populations for different SNP types5


https://doi.org/10.1101/023051

Figure S5: Manhattan plot of the Bayenv results.

all SNPs Z top 1% fixed

INTERGENICINTRONUTR_3_PRIMEUTR_5_PRIMESYNONYMOUS_STOPSYNONYMOUS_CODINGNON_SYNONYMOUS_CODINGSTOP_GAINEDSTOP_LOSTSTART_LOST

Prop

ortio

n of

SN

P ty

pes

0.0

0.2

0.4

0.6

0.8

1.0

Figure S6: Relative proportion of the different SNP types for all SNPs and three differentcandidate SNP sets. (Bayenv2.0 top 1% and completely fixed between high and low pops)

6


https://doi.org/10.1101/023051

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●● ●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●● ●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●● ●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●● ●●●●● ●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

Position [bp]

geno

me−

wid

e ra

nk (

Z)

●●●●●

0.0

0.2

0.4

0.6

0.8

1.0

7050000 7060000 7070000 7080000 7090000 7100000

AAT3G20170

AT3G20180

AT3G20190AT3G20200

AT3G20210

AT3G20220AT3G20230

AT3G20240

AT3G20250AT3G20260

AT3G20270

AT3G20280AT3G20290

AT3G20300

AT3G20310AT3G20320

AT3G20330

AT3G20340AT3G20350

AT3G20360

AT3G20362AT3G20365

AT3G20370

AT3G20380AT3G20390

AT3G20395

AT3G20400


Position [bp]

Alti

tude

[m]

7050000 7060000 7070000 7080000 7090000 7100000

617

877

990

2260

2282

B ●

●

●

●

●

●

●

●

UTR_5_PRIMENON_SYNONYMOUS_CODING

INTRONSYNONYMOUS_CODING

UTR_3_PRIMEINTERGENIC

STOP_GAINEDSTART_LOST

●●●●●●●● ●● ●●●●●●● ●●● ●●●●●●●● ●●● ●●●●● ●● ●●● ●●● ● ●●● ●●●●●●● ●● ●● ●●● ● ●●●●●● ●●●● ● ●● ●● ●●●●●●●●●●●●●● ●● ●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ● ●●● ●● ●● ● ●● ●●● ● ●●● ● ● ●●● ●●●●●●●● ●●●●●●●●●●●●●●●●●●● ●● ●●●● ●●●● ●●●● ●●●●●●● ●●●● ● ●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●● ●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●● ●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●● ●●●●●●●●● ●●●●● ●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●● ●●●●● ●●●●●●●●●●●●●●● ●●● ● ● ● ●●●●●●●● ●●● ●●●● ●● ●●● ● ●● ●●●●●●●●●● ● ●●● ●●●●●●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●● ●●●●●●●●●●●● ● ●●●●●●●●●●● ●● ● ●● ● ●● ● ●●●●● ●●●●●●●●●● ●●●●●●●●●●●● ●●●●●●● ●● ●● ●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●●●●●● ●●●●● ●●● ●●● ●●●●●●●●●●●●●●●●●●●●●●●●

Position [bp]

Cov

erag

e pe

r po

ol

7050000 7060000 7070000 7080000 7090000 7100000

010

2030

4050

6070

CJuvalLaatsch

FinailTerz

Vioz/Coro

Figure S7: Example of the candidate gene AtEHD1 (AT3G20290) in which allele frequencieswere highly differentiated between high and low altitude populations. (A) Plot of the genome-wide relative rank of the Z statistics. High values indicate a strong differentiation. (B) Plot ofthe relative allele frequencies of the two alleles segregating at a polymorphic site. The heightof the bar is proportional to the allele frequency in the pooled sequence data. (C) Sequencecoverage of the genomic region in the five different populations.

7


https://doi.org/10.1101/023051

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●● ●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●● ●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●

●●●●●

●●●●●●●●●●

●●●●●●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●

●●●●●●●●●●●●●●●

●●●●●

●●●●●

●●●●●

Position [bp]

geno

me−

wid

e ra

nk (

Z)

●●●●●

0.0

0.2

0.4

0.6

0.8

1.0

2230000 2240000 2250000 2260000 2270000 2280000

AAT3G07020

AT3G07025

AT3G07030AT3G07040

AT3G07050

AT3G07055AT3G07060

AT3G07070

AT3G07080AT3G07090

AT3G07100

AT3G07110AT3G07115

AT3G07120

AT3G07130AT3G07140

AT3G07150

AT3G07160AT3G07170

AT3G07180

AT3G07185AT3G07190

AT3G07195

AT3G07200AT3G07210


Position [bp]

Alti

tude

[m]

2230000 2240000 2250000 2260000 2270000 2280000

617

877

990

2260

2282

B ●

●

●

●

●

●

●NON_SYNONYMOUS_CODINGSYNONYMOUS_CODING

STOP_GAINEDINTERGENIC

INTRONUTR_5_PRIME

UTR_3_PRIME

●●●●●●●●●●●●●● ●●●●●●●●●●●● ●●●●●●●● ●●●●●●● ●●●●●●●●●●● ●●●●● ●● ●●● ● ●●●● ● ●● ●● ●●● ●●●● ●●●●●●●●●●●●●●●●●●● ●● ●●●● ● ●●● ● ● ● ● ●● ● ● ● ● ●● ● ●●●●●●● ●●●●●●●● ●●●● ●●● ●●●●●●●●●●● ●●●●●●● ●●●●●●● ●●●●●●●●●●●●●● ●●● ●●●●●●●●●●●●●●● ●●●●●●●●●●● ● ●● ● ● ●●●●●●● ● ● ●●●●●●●●●●●●●●●●●●●● ● ● ● ●● ● ● ● ● ●●●●● ● ●● ●● ● ● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●● ●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●● ● ●●●●●●●●●● ●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●● ● ●●●●●●●●●●●●● ●●●●●●●●●●● ●● ● ●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●● ● ●●●●●●●●● ●●●●●●●● ●●●● ●●●●●● ● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●

Position [bp]

Cov

erag

e pe

r po

ol

2230000 2240000 2250000 2260000 2270000 2280000

010

2030

4050

6070

CJuvalLaatsch

FinailTerz

Vioz/Coro

Figure S8: Example of a candidate gene (AT3G07130, encodes the purple acid phosphatase 15)in which allele frequencies differ strongly between the population at the highest altitude and thefour lower altitude populations. (A) Plot of the genome-wide relative rank of the Z statistics.High values indicate a strong differentiation. (B) Plot of the relative allele frequencies of thetwo alleles segregating at a polymorphic site. The height of the par is proportional to the allelefrequency. (C) Sequence coverage of the genomic region in the five different populations.

8


https://doi.org/10.1101/023051

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9


Rat

io d

elet

erio

us/to

lera

ted

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4 Juval

LaatschFinail

TerzVioz/Coro

*****

*

**

**

****

*****

Figure S9: Ratio of relative frequencies of deleterious and tolerated amino acid polymorphismsper frequency bin. If both polymorphism types were neutral, a ratio of 1 (dashed line) wouldbe expected for all frequency classes. Asterisk above bars represent ratios significantly differentfrom 1 (nominal p-value threshold 0.01, tested with a χ2-test).

9


https://doi.org/10.1101/023051

Mean Tajima D

Fre

quen

cy

−1.0 −0.5 0.0 0.5

050

100

150

200

A

Mean Tajima D

Fre

quen

cy

−0.5 0.0 0.5 1.0

050

100

150

200

250

B

Mean Tajima D

Fre

quen

cy

−1.5 −1.0 −0.5 0.0 0.5

020

4060

8010

0

C

Mean Tajima D

Fre

quen

cy

−1.5 −1.0 −0.5 0.0 0.5

020

4060

8010

0

D

Figure S10: Posterior distribution of Tajima’s D values in the high altitude population for thebest four ABC models. (A) Juval-Finail stepgrowth model. (B) Terz-Vioz/Coro exponentialgrowth model. (C) Juval-Finail exponential growth with migration model. (D) Terz-Vioz/Coroexponential growth with migration model. Observed values are indicated by the red verticallines.

10


https://doi.org/10.1101/023051

Supplementary Tables27

Table S1: Analysis of deviance table of a binomial GLM for dead biomass in response to differentdosages of UV-B.

Variable F-Valuedf/df P-Value

population 754/191 <0.001

UV-B dosage 8741/190 <0.001

population × UV-B dosage 224/186 <0.001

11


https://doi.org/10.1101/023051

Table S2: Average pairwise FST (lower triangle) and pairwise correlations of Bayenv2.0 corre-lation matrix among allele frequencies (upper triangle).

Juval Laatsch Finail Terz Vioz/Coro

Juval - 0.79 0.81 0.75 0.67Laatsch 0.12 - 0.77 0.99 0.71Finail 0.12 0.12 - 0.71 0.81Terz 0.12 0.12 0.14 - 0.62Vioz/Coro 0.14 0.12 0.12 0.16 -

12


https://doi.org/10.1101/023051

Tab

leS

3:P

rior

dis

trib

uti

ons

ofth

eJu

val-

Fin

ail

step

-gro

wth

and

Ter

z-V

ioz/

Coro

exp

on

enti

al

gro

wth

mod

els.

J-F

Ste

pgr

owth

J-F

Exp

onen

tial

mig

.T

-V/C

Exp

onen

tial

T-V

/C

Exp

on

enti

al

mig

.P

aram

eter

Min

Max

Min

Max

Min

Max

Min

Max

θa2.

811

.22.

811

.22.

811

.22.8

11.2

ρb

0.25

1.02

0.25

1.02

0.25

1.02

0.25

1.0

2N

high

0.00

11.

00.

001

1.0

0.00

11.

00.0

01

1.0

Nc high,growth

--

0.0

159.

7575

0.0

159.

7575

0.0

159.7

575

Nd high,growth,tim

e-

-0.

0118

750.

4975

0.01

1875

0.49

750.

011875

0.4

975

Tsp

lit

0.01

250.

50.

0125

0.5

0.01

250.

50.

0125

0.5

Me 12

--

0.0

10.0

--

0.0

10.0

M21

--

0.0

10.0

--

0.0

10.0

aT

het

ais

calc

ula

ted

as4∗N

0∗µ∗L

;N

0=

[50,

000:

200,

000]

;µ

=7∗

10−9;

L=

2000

(len

gth

of

sim

ula

ted

segm

ents

).bR

ho

isca

lcu

late

das

4∗N

0∗r∗L

;N

0=

[50,

000:

200,

000]

;r=

0.62

5∗

10−9;

L=

2000

(len

gth

of

sim

ula

ted

segm

ents

).cG

row

thra

teof

the

hig

hp

opu

lati

on.

dT

ime

ofth

eh

igh

pop

ula

tion

grow

th.

eM

igra

tion

rate

isca

lcu

late

das

4∗N

0∗m

.

13


https://doi.org/10.1101/023051

Tab

leS4:

Pos

teri

orm

od

elp

rob

abil

itie

sof

each

dem

ogra

ph

icsc

enar

ioca

lcu

late

dby

the

AB

Cpostpr

fun

ctio

nu

sin

gth

emnlogistic

met

hod

wit

ha

0.00

5to

lera

nce

tose

lect

the

bes

tgr

owth

mod

elfo

rth

eJu

val-

Fin

ail

and

Ter

z-V

ioz/

Coro

pop

ula

tion

pair

s.

Wit

hou

tm

igra

tion

Wit

hm

igra

tion

Pop

ula

tion

pai

rN

ogr

owth

Ste

pL

inea

rE

xp

onen

tial

No

grow

thS

tep

Lin

ear

Exp

on

enti

al

Ju

val-

Fin

ail

0.00

00.

000.

000

0.00

00.

000

0.00

10.0

00

0.9

99

Ter

z-V

ioz/

Cor

o0.

000

0.00

00.

000

0.00

00.

000

0.00

00.

000

1.0

00

14


https://doi.org/10.1101/023051

Table S5: Soil parameters of surface soil (5-10 cm) in the five population sites.

Population pH Organic matter Phosphate Potassium Magnesium% mg/100g mg/100g mg/100g

Juval 6.1 6.17 4.2 21 21Terz 6.2 4.9 5.5 29 26Laatsch 6.3 3.47 <1.0 8 8.3Finail 4.6 17.2 16 58 36Vioz/Coro 5.2 15.8 22 24 23

15


https://doi.org/10.1101/023051

Tab

leS

6:A

ll25

SN

Ps

wit

hZ

=0.

5in

the

Bay

env2

an

aly

sis.

Chromosome

bp

AGIID

SNP

type

Description

19,021,128

AT1G26090

Intron

P-loopcontainingnucleosidetriphosphate

hydrolasessuperfamilyprotein

111,056,544

AT1G31010

Synonymouscoding

Organellarsingle-stranded

DNA

bindingprotein

4(O

SB4)

2902,758

AT2G03060

Intron

AGAMOUS-L

IKE

30;Pollen

development,

pollen

exineform

ation

212,412,973

Intergen

ic2

13,606,764

AT2G31970

Non-synonymouscoding

Encodes

theArabidopsisRAD50homologue;

DNA

repair,RNA

processing

214,276,102

AT2G33760

Synonymouscoding

Pentatricopep

tiderepeat(P

PR)superfamilyprotein

214,376,598

AT2G34030

Intron

Calcium-bindingEF-handfamilyprotein

32,256,537

AT3G07130


PURPLE

ACID

PHOSPHATASE

15;Pollen

germination,seed

germination

33,897,644

AT3G12220

Intron

SERIN

ECARBOXYPEPTID

ASE-L

IKE

16;Proteolysis

34,298,285

AT3G13290


Varicose-related(V

CR)

34,427,242

AT3G13560


O-G

lycosylhydrolasesfamily17protein;Carb

ohydrate

metabolicprocess

34,878,614

Intergen

ic3

7,077,355

AT3G20290

Intron

ArabidopsisEps15homologydomain

proteinsinvolved

inen

docytosis(A

tEHD1)

318,716,502

AT3G50430


Unknow

nprotein

322,868,343

AT3G61780


Embryodefective1703(emb1703);

Embryodevelopmenten

dingin

seed

dorm

ancy

410,106,954

Intergen

ic4

13,574,591

AT4G27040

5’UTR

VPS22;Vesicle-m

ediatedtransp

ort

413,577,662

AT4G27050

Intron

F-box

/RNI-likesuperfamilyprotein

5824,047

AT5G03360

Intron

DC1domain-containingprotein;Oxidation-red

uctionprocess

55,587,807

AT5G17010

Intron

Majorfacilitatorsuperfamilyprotein;Transm

embranetransp

ort

55,903,750

AT5G17860


Calcium

exchanger

7(C

AX7)

56,252,514

Intergen

ic5

8,335,327

Intergen

ic5

20,429,419

Intergen

ic5

26,727,336

AT5G66930

Intron

Unknow

nprotein

16


https://doi.org/10.1101/023051

Table S7: Enriched GO Biological processes among the top Bayenv2.0 results with nominalp-values ≤ 0.05.

Significant genes All genes Description

3 5 Shade avoidance, Goslim: response to abiotic or biotic stimulus4 7 Secondary shoot formation, Goslim: developmental processes6 12 Leaf vascular tissue pattern formation, Goslim: developmental

processes19 71 Response to light stimulus, Goslim: response to abiotic or

biotic stimulus4 10 Negative regulation of transcription, Goslim: transcription4 12 Cuticle development, Goslim: developmental processes4 9 Heat acclimation, Goslim: response to stress14 57 Unidimensional cell growth, Goslim: other cellular processes5 12 Defense response signaling pathway, resistance gene-

dependent, Goslim: response to stress8 22 Chloroplast organization, Goslim: cell organization and biogenesis2 5 Flavonol biosynthetic process, Goslim: other cellular pro-

cesses11 42 Shoot development, Goslim: developmental processes4 6 Phenylpropanoid metabolic process, Goslim: other cellular

processes5 11 Phloem or xylem histogenesis, Goslim: developmental processes8 36 Flower development, Goslim: developmental processes3 7 Folic acid and derivative biosynthetic process, Goslim: other cel-

lular processes4 6 Ovule development, Goslim: developmental processes23 108 Multicellular organismal development, Goslim: developmental

processes6 24 Plant-type cell wall modification during multidimensional cell

growth, Goslim: other cellular processes

17


https://doi.org/10.1101/023051

Table S8: Enriched PO, KEGG, AraCyc categories among the top Bayenv2.0 results withnominal p-values ≤ 0.05.

Significant genes All genes Description

9 15 PO:0020144 expressed in apical meristem14 38 PO:0000229 expressed in flower meristem3 4 Superoxide radicals degradation6 12 Glucosinolate biosynthesis8 20 CDP-diacylglycerol biosynthesis I8 20 CDP-diacylglycerol biosynthesis II5 10 glycolipid desaturation5 11 PO:0020149 expressed in quiescent center4 9 Glucosinolate biosynthesis from phenylalanine3 6 Glucosinolate biosynthesis from tetrahomomethionine3 6 Glucosinolate biosynthesis from trihomomethionine5 11 Tetrahydrofolate biosynthesis II10 34 Protein export3 5 Mevalonate pathway I3 7 Glucosinolate biosynthesis from homomethionine3 7 Glucosinolate biosynthesis from hexahomomethionine3 7 Glucosinolate biosynthesis from pentahomomethionine2 3 Brassinosteroid biosynthesis II7 22 PO:0006085 expressed in root meristem2 2 PO:0007130 expressed during B reproductive growth3 7 Glucosinolate biosynthesis from dihomomethionine8 23 α-Linolenic acid metabolism3 4 Flavone and flavonol biosynthesis4 8 Isoleucine degradation I3 3 Anthocyanin biosynthesis (pelargonidin 3-O-glucoside, cyani-

din 3-O-glucoside)3 8 PO:0000372 expressed in metaxylem7 19 FormylTHF biosynthesis II

18


https://doi.org/10.1101/023051

Bayenv228

Table S9: GO terms among top Z signals with nominal p-value≤ 0.05 obtained from an analysiswith Gowinda.

Cat FDR Genes found All genes Description

CC 0.2918 29 125 GO:0009579 thylakoid, Goslim:other intracellu-

lar components

MF 0.30478875 3 4 GO:0004784 superoxide dismutase activity,

GOslim other enzyme activity

MF 0.30478875 7 14 GO:0008794 arsenate reductase (glutaredoxin)

activity, GOslim other enzyme activity

BP 0.30478875 2 2 GO:0016579 protein deubiquitination,

Goslim:protein metabolism

BP 0.30478875 3 5 GO:0009641 shade avoidance, Goslim:response

to abiotic or biotic stimulus

CC 0.30478875 9 22 GO:0009295 nucleoid, Goslim:other intracellular

components

CC 0.30478875 5 10 GO:0005819 spindle, Goslim:other intracellular

components

MF 0.30478875 208 1185 GO:0003700 transcription factor activity,

GOslim transcription factor activity

BP 0.3569411111 4 7 GO:0010223 secondary shoot formation,

Goslim:developmental processes

CC 0.3764346154 6 15 GO:0019005 SCF ubiquitin ligase complex,

Goslim:other intracellular components

CC 0.3764346154 50 221 GO:0005783 endoplasmic reticulum, Goslim:ER

BP 0.3764346154 6 12 GO:0010305 leaf vascular tissue pattern forma-

tion, Goslim:developmental processes

MF 0.3764346154 17 58 GO:0046872 metal ion binding, GOslim other

binding

CC 0.3783357143 8 21 GO:0009524 phragmoplast, Goslim:other cyto-

plasmic components

CC 0.44608 3 5 GO:0009574 preprophase band, Goslim:other

cytoplasmic components

BP 0.493291875 19 71 GO:0009416 response to light stimulus,

Goslim:response to abiotic or biotic stimulus

MF 0.5177496 13 56 GO:0016563 transcription activator activity,

GOslim other molecular functions

BP 0.5177496 4 10 GO:0016481 negative regulation of transcrip-

tion, Goslim:transcription

19


https://doi.org/10.1101/023051

BP 0.5177496 4 12 GO:0042335 cuticle development,


BP 0.5177496 4 9 GO:0010286 heat acclimation, Goslim:response

to stress

BP 0.5177496 14 57 GO:0009826 unidimensional cell growth,

Goslim:other cellular processes

BP 0.5177496 5 12 GO:0009870 defense response signaling pathway,

resistance gene-dependent, Goslim:response to

stress

BP 0.5177496 8 22 GO:0009658 chloroplast organization,

Goslim:cell organization and biogenesis

MF 0.5177496 6 15 GO:0008017 microtubule binding, GOslim pro-

tein binding

BP 0.5177496 2 5 GO:0051555 flavonol biosynthetic process,


MF 0.5183137037 13 48 GO:0008415 acyltransferase activity, GOslim

transferase activity

BP 0.5183137037 11 42 GO:0048367 shoot development,


BP 0.5257717857 4 6 GO:0009698 phenylpropanoid metabolic pro-

cess, Goslim:other cellular processes

BP 0.5337251724 5 11 GO:0010087 phloem or xylem histogenesis,


BP 0.5614648649 8 36 GO:0009908 flower development,


MF 0.5614648649 5 10 GO:0016207 4-coumarate-CoA ligase activity,

GOslim other enzyme activity

BP 0.5614648649 3 7 GO:0009396 folic acid and derivative biosyn-

thetic process, Goslim:other cellular processes

MF 0.5614648649 5 12 GO:0016291 acyl-CoA thioesterase activity,

GOslim hydrolase activity

BP 0.5614648649 4 6 GO:0048481 ovule development,


BP 0.5614648649 23 108 GO:0007275 multicellular organismal develop-

ment, Goslim:developmental processes

CC 0.5614648649 148 835 GO:0005634 nucleus, Goslim:nucleus

BP 0.5614648649 6 24 GO:0009831 plant-type cell wall modifica-

tion during multidimensional cell growth,


CC 0.5720152632 2 4 GO:0016272 prefoldin complex, Goslim:cytosol

20


https://doi.org/10.1101/023051

MF 0.5730625641 46 205 GO:0005215 transporter activity, GOslim trans-

porter activity

Table S10: KEGG pathways, PO terms and Aracyc networks among top Z signals with nominalp-value≤ 0.05


PO 0.0746 9 15 PO:0020144 expressed in apical meristem

PO 0.133415 14 38 PO:0000229 expressed in flower meristem

ARACYC 0.1537485714 3 4 superoxide radicals degradation

KEGG 0.1537485714 6 12 Glucosinolate biosynthesis

ARACYC 0.1537485714 8 20 CDP-diacylglycerol biosynthesis I

ARACYC 0.1537485714 8 20 CDP-diacylglycerol biosynthesis II

ARACYC 0.1537485714 5 10 glycolipid desaturation

PO 0.1630625 5 11 PO:0020149 expressed in quiescent center

ARACYC 0.235128 4 9 glucosinolate biosynthesis from phenylalanine

ARACYC 0.2970092857 3 6 glucosinolate biosynthesis from tetrahomome-

thionine

ARACYC 0.2970092857 3 6 glucosinolate biosynthesis from trihomomethio-

nine

ARACYC 0.2970092857 5 11 tetrahydrofolate biosynthesis II

KEGG 0.2970092857 10 34 Protein export

ARACYC 0.324188 3 5 mevalonate pathway I

ARACYC 0.3547026316 3 7 glucosinolate biosynthesis from homomethionine

ARACYC 0.3547026316 3 7 glucosinolate biosynthesis from hexahomome-

thionine

ARACYC 0.3547026316 3 7 glucosinolate biosynthesis from pentahomome-

thionine

ARACYC 0.3547026316 2 3 brassinosteroid biosynthesis II

PO 0.3764761905 7 22 PO:0006085 expressed in root meristem

PO 0.3764761905 2 2 PO:0007130 expressed during B reproductive

growth

ARACYC 0.4155586957 3 7 glucosinolate biosynthesis from dihomomethion-

ine

KEGG 0.4155586957 8 23 alpha-Linolenic acid metabolism

KEGG 0.42463625 3 4 Flavone and flavonol biosynthesis

ARACYC 0.4529372 4 8 isoleucine degradation I

ARACYC 0.4808680769 3 3 anthocyanin biosynthesis (pelargonidin 3-O-

glucoside, cyanidin 3-O-glucoside)

PO 0.4842432609 3 8 PO:0000372 expressed in metaxylem

ARACYC 0.4842432609 7 19 formylTHF biosynthesis II

21


https://doi.org/10.1101/023051

Table S11: Literature collections among top Z signals with nominal p-value≤ 0.05


LIT 0.002275 47 159 SOM3-down (TableS4 PubmedID:18545680)

LIT 0.002275 10 14 Changes in the expression levels of genes tar-

geted by microRNA and trans-acting short in-

terfering RNA assayed on ATH1 microarrays

LIT 0.002275 359 1945 genes with lower expression levels in 3d old

RBRcs seedlings compared to 3d old wild type

seedlings, grown in the presence of 1% sucrose

LIT 0.002275 98 456 Down-regulated expressed in the obe1-1 obe2-2

when compared to wild type (Table S1 Pubme-

dID:19392692)

LIT 0.003648 24 86 up-regulated genes upon infection of Arabidop-

sis with R. fascians D188

LIT 0.0079742857 268 1459 Up-regulated wt vs ag mutant Stage12 (Table

S2 PubmedID:19385720)

LIT 0.0079742857 173 908 Down-regulated 2-fold in WT-Rc50/WT-D (Ta-

ble S1 PubmedID:19529817)

LIT 0.008375 113 561 ABA Up-regulated Genes (692)

LIT 0.008375 140 707 cluster4 Expression clusters of H3K27me3-

enriched genes by K-means clustering.

LIT 0.008375 122 626 Overview of the 810 highly-responsive genes.

Additional file 6 lists genes whose log2(ratio)

was greater than 1.585 or lesser than -1.585 (cor-

responding to 3-fold change) in at least one of

the MA/M, SA/M or S/M comparisons in the

CATMA array experiment.

LIT 0.0084681818 125 606 Induced RNAs to 718 LEC2 (Table S2 Pubme-

dID:16492731)

LIT 0.0094441667 96 474 Preferentially expressed in seed (Table S 7 Pub-

medID:18539592)

LIT 0.0138346154 234 1272 Up-regulated 2-fold in YHB-D /WT-D (Table


LIT 0.0139393333 323 1766 genes with higher expression levels in 3d old


seedlings, grown in the presence of 1% sucrose

LIT 0.0139393333 290 1588 genes with lower expression levels in 3d old


seedlings, grown in the absence of sucrose

22


https://doi.org/10.1101/023051

LIT 0.016344375 230 1275 Genes Deregulated in det3 under Restrictive

Conditions (Nitrate)

LIT 0.0182861111 270 1430 genes with higher expression levels in 3d old


seedlings, grown in the absence of sucrose

LIT 0.0182861111 133 679 Up-regulated genes Inferred from Ratios r4 (Ta-


LIT 0.0196794737 123 673 down-differentially regulated by Pieris brassi-

cae oviposition on wild-type, coi1-1, and sid2-1

plants



LIT 0.0262841667 18 61 Up-regulated (1.5-fold change, p = 0.01) by

silicon supply in Arabidopsis plants infected

with Erysiphe cichoracearum (Table S1 Pubme-

dID:17082308)

LIT 0.0262841667 417 2474 ¿2x, vs. Sav+resc., root+seedl.

LIT 0.0262841667 73 330 Down-regulated carpel-expressed genes that

were used to enhance the prediction of carpel-

specific transcripts by ag in the predictions of

array elements of the oligonucleotide array rep-

resenting floral organ-expressed transcripts are

summarized. (Table S12 PubmedID:15100403)

LIT 0.0262841667 199 1097 Down-regulated 2-fold in YHB-Rc50/WT-D

(Table S1 PubmedID:19529817)

LIT 0.0268456 99 489 Up-regulated in the stn7-1 psad1-1 mutant com-

pared to Col-0. (Table S5 PubmedID:19706797)

LIT 0.0269688462 30 119 Decreased genes in TOC1-ox compared with

WT (Table S1 PubmedID:19816401)

LIT 0.0269855172 23 86 Genes significantly changing in at least two of

three conditions: downregulated in jaw-D leaves

or apices, upregulated in rTCP4:GFP plants

(logit-T p¡0.05 and common variance ¿ 2 fold).

Promoters were arbitrarily defined as 800 to 10

bp from ATG.

LIT 0.0269855172 364 2116 Differentially expressed genes using a p-value ¡

0.05 which are early seed specific (ESS) in the

GHL Data. (Table S2 PubmedID:18923020)

LIT 0.0269855172 14 41 genes repressed in stop1-mutant, and down-

regulated or stable in WT with Al treatment

23


https://doi.org/10.1101/023051

LIT 0.0274623333 53 239 Up-regulated carpel-expressed genes that were

used to enhance the prediction of carpel-specific

transcripts by ap2 in the predictions of array

elements of the oligonucleotide array represent-

ing floral organ-expressed transcripts are sum-

marized. (Table S12 PubmedID:15100403)

LIT 0.0279632258 63 300 pklpkr2 UP List of genes deregulated in pkl and

pkl pkr2 roots







LIT 0.0319874286 190 1017 Up-regulated 2-fold in YHB-Rc50/WT-D (Table


LIT 0.0319874286 142 819 Fold induction relative to T0 of genes signifi-

cantly differentially expressed upon auxin treat-

ment

LIT 0.0319874286 277 1559 Genes differentially regulated in the SCR

microarray time course (Table S2 Pubme-

dID:20596025)

LIT 0.0382271795 22 84 100 galls distinctive DEG with the lowest Fc List

of the co-expresed DEG 3 days after infection

of Meloidogyne javanica(MJ) between GCs and

galls and the galls distinctive genes

LIT 0.0382271795 108 575 Down-regulated genes Inferred from Ratios r5


LIT 0.0382271795 7 16 dn,Expression of antioxidant network genes fol-

lowing the transition to darkness

LIT 0.0382271795 8 17 dn,aox1a treated vs aox1a normal,Relative tran-

script abundance for genes encoding anti-

oxidant defence components located in Ara-

bidopsis

LIT 0.03849925 91 439 dn,Microarray analysis of Wild type vs

35S:ZAT10 line 14 (OE)

LIT 0.0430307317 226 1270 Up-regulated wt vs ag mutant after bolting (Ta-


LIT 0.0448921429 29 123 Increase in both 2HS IP and 9HS IP (Table S4

PubmedID:18665916)

24


https://doi.org/10.1101/023051

LIT 0.0493845455 199 1106 Down-regulated 2-fold in YHB-D /WT-D (Table


LIT 0.0493845455 57 267 DELLA-repressed (DELLA-down) genes in the

imbibed ga1-3 seeds

LIT 0.0511073333 27 101 UP enhanced regulation in control Genes dif-

ferentially expressed in response to drought in

35S:ABF3 and control plants

LIT 0.0539280435 61 298 DN differentially expressed genes in AtNudt7-1

versus Wild-type plants growing in 12:3:1 pot-

ting mix

LIT 0.0565732653 42 167 Dwon-regulated carpel-expressed genes that


specific transcripts by ap3 in the predictions of




LIT 0.0565732653 10 28 Down-regulated in sdg4 mutant flowers (Table 1

PubmedID:18252252)

LIT 0.0565732653 13 39 Up regulated (¿ 2 fold) with drought in WT and

C2 (Table S3 PubmedID:18849493)



LIT 0.0623896552 362 2064 Upregulated genes in hyl1-2. (Table S3 Pubme-

dID:19633021)

LIT 0.0623896552 35 155 Down-regulated in the stn7-1 mutant compared

to Col-0. (Table S3 PubmedID:19706797)

LIT 0.0623896552 117 592 Nuclear RP Up Unambig (Table S2 Pubme-

dID:20675573)

LIT 0.0623896552 185 1031 Downregulated genes between TEV and TEV-at

infected plants (TableS3 PubmedID:18545680)



LIT 0.0623896552 125 648 EP-ESS generated from the list of differentiall

expressed genes using p -value ¡ 0.05. (Table S5

PubmedID:18923020)

LIT 0.0623896552 80 406 downregulated in the imbibed ga1-3 seeds (GA-

upregulated or GA-up)

LIT 0.0623896552 169 952 Down-regulated wt vs ag mutant after bolting


25


https://doi.org/10.1101/023051

LIT 0.0632477966 49 220 Down-regulated genes from meta-analysis

(q¡0.05) by GL3 DM 24hrs (Table S 2 Pubme-

dID:19247443)

LIT 0.0637395 28 118 Decrease of effect of pkl and of uniconazole on

transcript level of gene. (Table S 5 Pubme-

dID:18539592)

LIT 0.063917541 117 611 significantly differential expression patterns be-

tween wild-type, mpk4 and 35S-MKS1 plants

LIT 0.0705620968 6 12 dn,Expression of antioxidant network genes that

do not have a clear cellular localization following

transition to darkness



transcripts by pi in the predictions of array el-

ements of the oligonucleotide array represent-



LIT 0.0713777273 88 452 Up-regulated in brm-5/essp3 Leaves (Table S1

PubmedID:18508955)

LIT 0.0713777273 12 40 Up-regulated genes (top 50)in the albino mutant


LIT 0.0713777273 352 2063 Venn analysis of differentially expressed genes;

down-regulated genes (TableS4b Pubme-

dID:18400103)

LIT 0.0718035211 211 1184 Down-Genes differentially regulated in dms4

compared to drd1. (Table S3 Pubme-

dID:20010803)

LIT 0.0718035211 300 1745 dn,Significant genes changing during CaLCuV

infection

LIT 0.0718035211 10 30 Repressed SSTF Class 3 genes that are Re-

pressed in WT-R1, WT-Rc and/or pifq-D rela-

tive to WT-D. (Dataset4 PubmedID:19920208)

LIT 0.0718035211 16 67 Upregulated -AGL24 -genes by microarray anal-

ysis (Table S1 PubmedID:18339670)

LIT 0.0718035211 77 382 UP Differentially regulated genes in hsp70-15

knockout plants.

LIT 0.0718127778 9 26 Downregulated in shoot between [NH4NO3

+CO(NH2)2]- and [NH4NO3]-Arabidopsis

grown plants (reference). (TableS5 Pubme-

dID:18508958)

26


https://doi.org/10.1101/023051

LIT 0.0720809091 115 619 Up-regulated 2-fold in WT-Rc15/WT-D (Table


LIT 0.0720809091 146 851 Down-regulated by WT -N vs. WT FN (Dataset

2 PubmedID:19933203)

LIT 0.0720809091 122 618 Up-6x2-List of genes that are upregulated or

downregulated in both the platforms. (Table S3

PubmedID:20406451)

LIT 0.0720809091 335 2016 Down-regulated expressed genes of Agrobac-

terium tumefaciens(P¡0.01 all)-induced Ara-

bidopsis tumors versus tumor-free inflorescence

stalk tissue. (Table S3 PubmedID:17172353)

LIT 0.0720809091 146 852 Up-regulated by WT FN vs. WT -N (Dataset 1

PubmedID:19933203)

LIT 0.0830333333 124 674 Up-regulated at least 2-fold after a shift from

growth light (100 mol photons m-2 s-1) to 3 h

high fluence blue light (200 mol photons m-2 s-1)


LIT 0.0853624051 10 31 LED 7 (up; all tissues) in root tissue cluster

(Arabidopsis) (TableS4 PubmedID:15918878)

LIT 0.086574875 21 98 up-regulated by the indole-3-acetic acid treat-

ment.

LIT 0.0872672289 8 24 up,aox1a normal vs Col-0 normal,Relative tran-



bidopsis

LIT 0.0872672289 11 33 down-Genes responsive to WUS induction

LIT 0.0872672289 30 130 Down-regulated genes in uninfected gh3.5-1D

versus wild-type plants (Table S2 Pubme-

dID:17704230)

LIT 0.0912563529 23 97 genes that are induced in Ler but not hy5-1:

LIT 0.0912563529 351 2112 Down-regulated in Agrobacterium tumefaciens-

induced tumors of Arabidopsis thaliana. (Table


LIT 0.0928162791 56 280 Up-regulated at least 2-fold in flu versus wild

type, ex1/flu versus wild type, ex2/flu versus

wild type, and ex1/ex2/flu versus wild type

(Data Set 1 PubmedID:17540731)

LIT 0.0930332184 16 64 dn,Genes differentially expressed in

P35S:PAR1-GG seedlings

27


https://doi.org/10.1101/023051

LIT 0.09601875 145 781 Up-regulated 2-fold in WT-Rc50/WT-D (Table


LIT 0.0972077528 89 461 Induced by auxin during lateral root initiation

in tissues surrounding the initiation site (Table


LIT 0.1028495604 124 662 Down-differentially regulated in WS-2 compared

to Col-O. (Table S9 PubmedID:19706797)

LIT 0.1028495604 8 20 Upregulated commonly in the gra-D mutants

and located outside of the duplications. (Table


LIT 0.1089384783 40 170 Up-genes were selected if they were detected

as differentially expressed upon both the Dex

and Dex+Chx treatments but not upon the Chx

treatment (Table S5 PubmedID:20360106)

LIT 0.1124013978 44 208 genes that are induced in Col but not cop1-4:

LIT 0.1124185106 20 75 Up-regulated-WT-temperature(Sample 29C /

Sample 20C) (Table S1 PubmedID:19686536)

LIT 0.1176825263 262 1557 Genes Deregulated in det3 under Restrictive

Conditions (Temperature)

LIT 0.11889 5 12 TF more induced-Differential gene expression

analysis of hypoxic stress AND cell types

(DatasetsS7 PubmedID:19843695)

LIT 0.1219590722 9 27 Down- regulated Secretory Pathway Genes

(SPGs) with modulated levels of transcripts in

5- and 11-day-old hypocotyls (File 4 Pubme-

dID:19878582)

LIT 0.1265104902 37 178 Genes Induced Maximally by Glucose at 2 Hours

(218)

LIT 0.1265104902 95 527 Genes Deregulated in det3 under Permissive

Conditions

LIT 0.1265104902 355 2113 OST generated from the list of differentiall ex-

pressed genes using p -value ¡ 0.05. (Table S5

PubmedID:18923020)

LIT 0.1265104902 7 20 Down-expression levels of glucosinolate-

metabolism and primary sulfur-metabolism

genes in response to MP (Table 3 Pubme-

dID:17189325)

LIT 0.1265104902 81 413 Up-regulated by pnp1-1 +P vs WT +P in three

hours (Table S3 PubmedID:19710229)

28


https://doi.org/10.1101/023051

LIT 0.1278679612 5 9 Significantly co-up-regulated in dcl1-7, dcl4-2,

and rdr6-15 (Table S 1 PubmedID:16129836)

LIT 0.1289255238 51 246 Repressed expressed genes upon SA treatment.

(TableS 1 PubmedID:17496105)

LIT 0.1289255238 64 326 Up-fis1x2-List of genes that are upregulated or


PubmedID:20406451)

LIT 0.1291354717 120 681 Down-regulated 2-fold in WT-Rc15/WT-D (Ta-


LIT 0.1294504673 31 141 Induced-3h-NO3 (Table S 6 Pubme-

dID:16299223)

LIT 0.1295172222 31 160 Affymetrix expression data validated by qRT-

PCR

LIT 0.1296294737 16 59 Diff-The number of unique hits per million are

shown for Arabidopsis genes greater than 1000

bases long in wild type shoots. A value of 0

indicates that, at the level of sequencing depth

used in this work, no reads aligning uniquely to

that particular transcript were found. (Table S2

PubmedID:20512117)

LIT 0.1296294737 4 10 Induced-group in Slightly later in the diurnal cy-

cle ¿2-fold changes within 30 min after adding

15 mM Suc to C-starved seedlings was extracted

from Osuna et al. (2007). (Table II Pubme-

dID:18305208)

LIT 0.1296294737 200 1174 The organ-expressed genes identified with the

oligonucleotide array are listed (Table S7 Pub-

medID:15100403)

LIT 0.1296294737 71 374 Microarray analysis of 35S:DREB2A CA

plants. Genes with fold change of ¿2 are

listed.–fluorescence intensity of each cDNA of

35S:DREB2A CA/fluorescence intensity of each

cDNA of vector control line.

LIT 0.1296294737 389 2267 Proteins identified by MS/MS in isolated wt and

ppi2 plastids

LIT 0.1296294737 16 59 Up-regulated 24 hours after infiltration of M.

persicae saliva and their expression frofile af-

ter M. persicae feeding. (Table S1 Pubme-

dID:19558622)

29


https://doi.org/10.1101/023051

LIT 0.1300405085 5 14 Down-regulated genes of the GO category

response-to-salt-stress in hmgb1 (Table S1 Pub-

medID:18822296)

LIT 0.1300405085 22 103 Upregulated genes-Overlapping between the

present study and that of Himanen et (2004)

with the cluster combination and the cluster

number and profile, respectively (Table S 4 Pub-

medID:16243906)

LIT 0.1300405085 70 372 Upregulated in BRZ Col (Table S4 Pubme-

dID:18599455)

LIT 0.1300405085 23 102 IAA up-regulated genes induction not affected

2 fold or more up or down by by simultaneous

glucose treatment

LIT 0.13463675 19 86 ARR1SR BA15 vs ARR1SR BA0 down,Lists of

all significantly regulated genes in the wild type

and line ARR1-S-8

LIT 0.13463675 53 263 Down-regulated genes -spl/nzz (TableS1 Pub-

medID:17905860)

LIT 0.1411063636 19 77 pklpkr2 UP H3K27me3 List of genes deregu-

lated in pkl and pkl pkr2 roots

LIT 0.1429948361 245 1424 differently regulated –cell cycle regulation dur-

ing the growth process of leaves 1 and 2 of Ara-

bidopsis

LIT 0.1433647967 20 78 Genes suggested to be involved in the transition

from a vegetative state to flowering and their

predicted regulators

LIT 0.1447189516 55 260 Depressed by KIN10 markedly overlaps with

that induced by starvation conditions and is an-

tagonized by increased sugar availability (600

genes). (Table S4 PubmedID:17671505)

LIT 0.1495619841 40 186 Increased genes in TOC1-ox compared with WT


LIT 0.1495619841 42 202 DELLA-upregulated (DELLA-up) genes in the

ga1-3 young flower buds

LIT 0.14981 12 46 Down-regulated genotype (elf3) x temperature

of 23 interaction (p¡0.01, uncorrected) (Table S4

PubmedID:19187043)

30


https://doi.org/10.1101/023051

LIT 0.1501121875 202 1194 Down-regulated expressed genes of Agrobac-

terium tumefaciens (P¡0.01 + functional

categories)-induced Arabidopsis tumors versus

tumor-free inflorescence stalk tissue. (Table S3

PubmedID:17172353)

LIT 0.1509633333 7 22 dn,aox1a treated vs Col-0 treated,Relative tran-



bidopsis

LIT 0.1509633333 158 950 Downregulated genes in hyl1-2. (Table S5 Pub-

medID:19633021)

LIT 0.1509633333 45 221 Genes Repressed maximally by Glucose at

6Hours (275)

LIT 0.1509633333 90 480 ARR1SR vs Col-0 down,Lists of all significantly

regulated genes in the wild type and line ARR1-

S-8

LIT 0.1512332331 127 732 Downregulated Se-reponsive genes in root tissue

(Table 1A PubmedID:18251864)

LIT 0.1527363433 36 161 Repressed-4h-carbon fixcation (Table S 6 Pub-

medID:16299223)

LIT 0.1554525899 10 30 Down-Coregulated Genes in stn7-1, psad1-

1, and psae1-3 Leaves (Table 3 Pubme-

dID:19706797)

LIT 0.1554525899 6 20 common ABA-responsive genes that reduced in

both srk2d/e/i and areb1/areb2/abf3 triple mu-

tants in comparison to WT after 6 h of treat-

ment with 50 ?M ABA

LIT 0.1554525899 11 43 dn,Diurnal gating of cold-responsive TFs.the

difference between the expression attained af-

ter morning cold treatment at ZT2 (Cm) versus

evening cold treatment at ZT14 (Ce)

LIT 0.1554525899 50 260 Genes Induced Maximally by Glucose at 6 Hours

(343)

LIT 0.1554525899 8 23 Down regulated at least 4 times in jaw-1D plants

(Table 2 PubmedID:12931144)

LIT 0.1564065 331 2005 Up-Genes differentially regulated in dms4 com-

pared to drd1. (Table S3 PubmedID:20010803)

LIT 0.1564184507 76 398 Genes showing more than twofold reduced

transcript level in 35S:AtMYB44 Arabidopsis

treated with 250 mM NaCl

31


https://doi.org/10.1101/023051

LIT 0.1564184507 57 304 down-regulated in 14 day-old leaves of Ara-

bidopsis haf2-1 mutant as assayed by CATMA

microarray

LIT 0.1584254545 69 374 Light up-regulated in Root (TableS6 Pubme-

dID:15888681)

LIT 0.167433125 13 55 Singificance Up¿4.0FC (S 02 Pubme-

dID:15668208)

LIT 0.1676214857 40 219 Up-regulated by AZC (Dataset1 Pubme-

dID:19244141)

LIT 0.1676214857 27 131 Genesrepressed (DOWN) in leaves of long-term

experiment : raw and normalized values and

flags are reported. Results for the three repli-

cates are reported.

LIT 0.1676214857 103 588 up,genes that are consistently affected in triple

mutant pollen

LIT 0.1676214857 23 107 Suppressed by ABA in rop10-1 (Table S 4 Pub-

medID:16258012)

LIT 0.1676214857 85 465 Down-expressed after SLWF nymph feeding


LIT 0.1676214857 8 30 Genes with specific expression in light-treated

seedlings.

LIT 0.1676214857 26 124 Genes Positively Regulated by HY5 (Table 1

PubmedID:17337630)

LIT 0.1676214857 33 161 Repressed-3h-mannitol (Table S 6 Pubme-

dID:16299223)

LIT 0.1676214857 24 126 List of genes up-regulated 2 h after treatment

with IAA

LIT 0.1676214857 17 72 up,Microarray identification of genes differen-

tially expressed in wrky33-1 mutants as com-

pared to wild-type

LIT 0.1676214857 102 552 Down-regulated genes -ap3es (TableS1 Pubme-

dID:17905860)

LIT 0.1676214857 69 368 OST-ESS generated from the list of differentiall

expressed genes using p -value ¡ 0.05. (Table S5

PubmedID:18923020)

LIT 0.1676214857 9 33 Down regulated (¿ 2 fold) with drought in WT

and C2 (Table S5 PubmedID:18849493)

LIT 0.1676214857 34 163 dn,Overview of MeJA-responsive genes(methyl

JA)

32


https://doi.org/10.1101/023051

LIT 0.1676214857 3 6 Repressed by infections of PPV and other pos-

itive sense RNA viruses (Table S8 Pubme-

dID:18613973)

LIT 0.1676214857 47 252 Down-regulated expressed in the obe1i obe2i

mutant seedlings when compared to wild type


LIT 0.1676214857 26 121 Up-regulated in 4hPT v 0.5hPT (Table S11

PubmedID:19714218)

LIT 0.1676214857 120 707 Genes differentially expressed in response to

Sclerotinia sclerotiorum(S-sclerotiorum) infec-

tion in leaves of in wild-type and coi1 mutant

LIT 0.1676214857 74 398 double/wt triple/wt DN genes significantly af-

fected in either the double(agl66 agl104-2),

triple(agl65 agl66 agl104-1), or quadruple(agl65

agl66 agl94 agl104-1) mutant pollen.

LIT 0.1676214857 29 134 UPR up-regulated genes dependent on bZIP28

function or expression

LIT 0.1676214857 104 595 Up-regulated expression (oler rapa FDR aver-

age CommVarsu) between the diploid progeni-

tors B. rapa and B. oleracea (Dataset S1 Pub-

medID:19274085)

LIT 0.1676214857 28 142 Up-regulated genes -35S::atNF-YB1 (TableS3

PubmedID:17923671)

LIT 0.1676214857 88 506 Down-regulated by COL vs atAF bgh (Table 2

PubmedID:18694460)

LIT 0.1676214857 4 9 Up-regulated specifically in flu after paraquat

treatmenta (Table S3 PubmedID:14508004)

LIT 0.1676214857 14 57 Co-expressed with higher Fc values in GCs than

in galls List of the co-expresed DEG 3 days after

infection of Meloidogyne javanica(MJ) between

GCs and galls and the galls distinctive genes

33


https://doi.org/10.1101/023051

LIT 0.1676214857 119 678 Up- regulated expressed Arabidopsis thaliana

(Ws-2) genes Suppl-Figure 2A: Table lists all

genes shown in Figure 2A differentially tran-

scribed in the following experiments: (i) Three

hours post inoculation of wounded inflorescence

stalks with A. tumefaciens strain C58 (C58 3

hpi) and (ii) with strain GV3101 (GV3101 3

hpi); (iii) six days post inoculation with strain

C58 (C58 6 dpi) and (iv) with GV3101 (GV3101

6 dpi), as well as (v) 35 day-old tumors induced

by strain C58 (tumor 35 dpi). (DatasetsS 1 Pub-

medID:19794116)

LIT 0.1676214857 2 3 Decreased in vte2 relative to wild-type at 3 days.


LIT 0.1676214857 64 382 down-Genes differentially regulated by Pieris

brassicae oviposition

LIT 0.1676214857 13 45 Induced transcripts(top 50) by ethylene (Table


LIT 0.1676214857 19 95 OtsB DN Genes more than 2-fold affected com-

pared with the wild type

LIT 0.1676214857 6 18 Up-expression levels of glucosinolate-

metabolism and primary sulfur-metabolism

genes in response to EC (Table 3 Pubme-

dID:17189325)

LIT 0.1685826136 143 841 UP WUSCHEL(WUS) REGULATION

SCORES for WUS response genes.

LIT 0.1732157303 56 277 Significantly regulated -uncapped predicted

miRNA (termed miSVM, Lindow et al., 2007)

targets. Only genes with uncapped/total

mRNA ratio ¿ 2 and a P value ¡ 0.01 in at least

one time point during early flower development

are shown. (Tables4 PubmedID:18952771)

LIT 0.1732157303 285 1674 Down-[ga1] vs [Col-0II] (Table S2 Pubme-

dID:20844019)

LIT 0.1759167222 92 514 Up-Further analysis indicated 640 genes in-

creased, their transcript levels by 4-fold or more

(p ? 0.001) as result of the dehydration stress.


34


https://doi.org/10.1101/023051

LIT 0.1759167222 44 232 Genes induced (UP) in the medium-term exper-

iment : raw and normalized values and flags are

reported. Results for the three replicates are re-

ported.

LIT 0.1779594737 45 216 Down-regulated in the stn7-1 psae1-3 mu-

tant compared to WS-2. (Table S7 Pubme-

dID:19706797)

LIT 0.1779594737 23 114 IAA up-regulated genes up-regulated by glucose

treatment alone.

LIT 0.1779594737 70 371 Opposite regulation (induced in BL/H3BO3,

repressed in GCs) co-expresed and distinctive

DEG between differentiating vascular cell sus-

pensions (BL/H3BO3) and GCs.

LIT 0.1779594737 37 187 Down-regulated in ahg2-1 (log2 ¡ -1 or log2 ¿1)


LIT 0.1779594737 28 133 pkl UP List of genes deregulated in pkl and pkl

pkr2 roots

LIT 0.1779594737 6 16 Differentially regulated -Targeted by NA small

RNAs (Table S1 PubmedID:17400893)

LIT 0.1779594737 7 20 Down-regulated and auxin-related genes in

tengu-transgenic plants as identified by microar-

ray analysis (Table 1 PubmedID:19329488)

LIT 0.1779594737 10 39 Reduced RNA level genes with at least a

5-fold in tdt-1 mutants (Table S1 Pubme-

dID:17513503)

LIT 0.1779594737 22 92 Down-regulated in the psae1-3 mutant com-

pared to WS-2 (Table S6 PubmedID:19706797)

LIT 0.1779594737 346 2108 EP generated from the list of differentiall ex-

pressed genes using p -value ¡ 0.05. (Table S5

PubmedID:18923020)

LIT 0.1779876963 36 167 ZAT12 dn,Overrepresentation of the CBF2 and

ZAT12 regulons among selected gene lists

LIT 0.1801860104 7 21 Up-miRNA targets significantly up-regulated

at least two-fold in dcl1 early globular em-

bryos relative to wild-type (Table S1 Pubme-

dID:21123653)

35


https://doi.org/10.1101/023051

LIT 0.1801860104 35 155 Down-regulated carpel-expressed genes that


specific transcripts by pi in the predictions of




LIT 0.1814145641 4 10 up,Complete list of genes diferentially expressed

in leaves of tpc1-2 versus WT

LIT 0.1814145641 23 116 down-regulated by the indole-3-acetic acid treat-

ment.

LIT 0.1827628141 102 584 Up-regulated in the psad1-1 mutant compared

to Col-0. (Table S4 PubmedID:19706797)

LIT 0.1827628141 2 3 Up regulated with drought (¿2 fold) in InsP 5-

ptase plants (Table S4 PubmedID:18849493)

LIT 0.1827628141 54 307 down-regulated drl1-2/Ler

LIT 0.1827628141 26 125 Down -regulated genes displaying differential

expression between mature and immature lat-

eral nectaries - MLN/ ILN . (Table S9 Pubme-

dID:19604393)

LIT 0.1838005366 15 60 Up- regulated expressed Arabidopsis thaliana

(Ws-2) genes Supplemental Figure 4 that are

affected either by C58 or GV3101 at 6hpi

(DatasetsS 1 PubmedID:19794116)

LIT 0.1838005366 34 165 N-terminal acetylated peptides(N-ace-peptid)

identified by MS/MS in Toc159cs leaves

LIT 0.1838005366 24 110 Aza-dC+TSA downregulated (TableS1 Pubme-

dID:15516340)

LIT 0.1838005366 7 29 Prominent upregulated genes in steroid-

inducible MKK3DD plants

LIT 0.1838005366 28 129 The 5% most highly expressed genes in mature

Arabidopsis trichomes

LIT 0.1838005366 7 22 Greatest up-regulated genes by 1.0 ?m IAA at

3h (Table 5a PubmedID:15086809)

LIT 0.184001699 9 35 Down Regulation of Cold/Light-Responsive

Genes at Least Two-Fold (P-value less than 0.05

at least in one condition)

LIT 0.184435782 5 15 Induced during PPV infection but either in-

duced or suppressed by infections of other pos-

itive sense RNA viruses (Table S8 Pubme-

dID:18613973)

36


https://doi.org/10.1101/023051

LIT 0.184435782 117 684 downregulated genes in the ga1-3 young flower

buds (GA-upregualted or up)

LIT 0.184435782 97 533 Down-regulated genes -ms1 (whole inflores-

cence) (TableS1 PubmedID:17905860)

LIT 0.184435782 57 312 Down-2x6-List of genes that are upregulated or


PubmedID:20406451)

LIT 0.184435782 54 289 DELLA-repressed (DELLA-down) geens in the

ga1-3 young flower buds

LIT 0.1864415566 55 311 down-regulated elo2/Ler

LIT 0.1883802347 60 313 upregulated genes in the ga1-3 young flower

buds (GA-downregulated or GA-down)

LIT 0.1911436449 39 221 up,Differentially expressed genes (548) that re-

sponded to 35 ?M NAE12:0 treatment in 4-day

old Arabidopsis seedlings

LIT 0.1916938889 23 120 down-regulated genes upon infection of Ara-

bidopsis with R. fascians D188

LIT 0.1916938889 10 38 Induced by both M. persicae feeding (De Vos et

al. 2005) and saliva infiltration (Table 1 Pub-

medID:19558622)

LIT 0.1919147465 6 16 Down-Genes affected in pye-1 mutants by Fe.–

Microarry analysis of wild-type and pye-1 mu-

tants after 24 hours of exposure to +Fe or Fe

media. (Table S3 PubmedID:20675571)

LIT 0.1941900889 24 113 Upregulated BBM target genes identified as

significantly in all three statistical analy-

ses (ANOVA, SAM, t-test) (TableS2 Pubme-

dID:18663586)

37


https://doi.org/10.1101/023051

LIT 0.1941900889 127 742 Differentially regulated by 1.5 fold (P ¡0.05) in

the wild-type plants by chitooctaose 30 min-

utes after treatment as revealed by the compari-

son between the chitooctaose-treated and water

treated samples. The expression pattern of these

genes in the similarly treated mutant plants was

also included after the wild-type Data. WT:

wild-type Col-0; Mu: the atLysM RLK1 mu-

tant; 8mer: treatment with chitooctaose; wa-

ter: treatment with distilled water; I, II, and

III: biological replicates. False Discovery Rate

(Number) of 200 permutations: Median: 0.0%

(0), and 90th percentile: 99.9% (889). (Table 1

PubmedID:18263776)

LIT 0.1941900889 3 8 Up-regulated on exhibit uniconazole-P-

dependent based on the criteria (Pkl vs.

Wt p ¡ 0.05 ?? Upkl vs. Uwt p ¡ 0.05)

and for which the corresponding transcript

level is elevated two-fold or more (Table S2

PubmedID:12834400)

LIT 0.1941900889 12 43 Up-regulated genes in MYB34 (Table S4 Pub-

medID:18829985)

LIT 0.1941900889 78 403 Decreased-regulation by KIN10 (1021 genes).


LIT 0.1941900889 78 403 Down-regulated-KIN10 (Table S6 Pubme-

dID:18305208)

LIT 0.1941900889 111 647 Jasmonate-esponsive genes (22hrs)in stamens

of Arabaidopsis thaliana (Table S1 Pubme-

dID:16805732)

LIT 0.1941900889 8 27 Repressed-group in Start of the extended night

¿2-fold changes within 30 min after adding 15

mM Suc to C-starved seedlings was extracted

from Osuna et al. (2007). (Table II Pubme-

dID:18305208)

LIT 0.1963785903 55 293 D - Genes repressed by Cd 15 and 30 M

(447),Genes significantly induced or repressed

exclusively by Cd2+ 30 M or by both Cd2+ 15

M and 30 M

38


https://doi.org/10.1101/023051

LIT 0.1963785903 119 661 Down-regulated in the stn7-1 psad1-1 mu-

tant compared to Col-0. (Table S5 Pubme-

dID:19706797)

LIT 0.2000454852 297 1832 Up-regulated expressed genes of Agrobacterium

tumefaciens(P¡0.01 all)-induced Arabidopsis tu-

mors versus tumor-free inflorescence stalk tissue.


LIT 0.2000454852 250 1511 Up-significantly altered expression (Dark Fold

Change) in 35S:MIF1 seedlings (false discovery

rate ¡0.1; MIF1, MINI ZINC FINGER 1) (Table


LIT 0.2000454852 68 372 DN Differentially regulated genes in hsp70-15

knockout plants.

LIT 0.2000454852 45 233 Down-regulated Gene List-Cold (Table S 6 Pub-

medID:16214899)

LIT 0.2000454852 45 235 Up-2x6-List of genes that are upregulated or


PubmedID:20406451)

LIT 0.2000454852 121 681 UP msh1-recA3 transcripts specifically regu-

lated by a certain mitochondrial impairment.

LIT 0.2000454852 133 778 Up-regulated at least 2-fold after a shift from

growth light (100 mol photons m-2 s-1) to 3 h

1000 mol photons m-2 s-1 in wild type (Table


LIT 0.2000454852 78 440 DN msh1-recA3 transcripts specifically regu-

lated by a certain mitochondrial impairment.

LIT 0.2000454852 11 45 Negative expression of list of non-redundant

genes that were differentially expressed compar-

ing flowering mutants to their near isogenic con-

trols (TableS2 PubmedID:15908439)

LIT 0.2000454852 31 157 Significant down-regulation in 72hrs in re-

sponse to iron-deprivation (Table S2 Pubme-

dID:18436742)

LIT 0.2009856574 23 101 up,Genes showing significant change in expres-

sion level between wild type and transgenic lines

LIT 0.2009856574 24 108 Cold Down-regulated Genes Whose Expres-

sion is Affected by ice1 (Table S 13 Pubme-

dID:16214899)

LIT 0.2009856574 81 459 upregulated in the imbibed ga1-3 seeds (GA-

downregulated or GA-down)

39


https://doi.org/10.1101/023051

LIT 0.2009856574 9 37 Commonly down regulated (¿ 2 fold) with

drought (161 transcripts) (Table S8 Pubme-

dID:18849493)

LIT 0.2009856574 6 20 Induced columella markers at regeneration 22

hrs compared to regeneration 0 hrs (Table S2

PubmedID:19182776)

LIT 0.2009856574 11 39 Up-regulated in CandiDate BABY BOOM tar-

get genes in DEX + CHX-treated 35S::BBM:GR

seedlings as compared to DEX + CHX-

treated wild-type seedlings (Table2 Pubme-

dID:18663586)

LIT 0.2009856574 19 88 All Genes That Pass Significance Criteria fol-

lowing Induction of GLK1.

LIT 0.2009856574 61 325 Up-regulated genes that were nonadditively ex-

pressed in allopolyploid line 1200 relative to the

1:1 parent mix–parentmix 5250 FDR Common-

Var (Dataset S5 PubmedID:19274085)

LIT 0.2009856574 7 22 Down-regulated by Sulfur and SLIM1 (Table S1

PubmedID:17114350)

LIT 0.2009856574 3 8 up,Genes up-regulated and down-regulated in 9-

h dehydrated ahk1 mutant

LIT 0.2009856574 9 32 Late repressed TFs in Jasmonate respon-

sive transcription factors (Table S3 Pubme-

dID:16805732)

LIT 0.2009856574 136 783 Down-regulated genes Inferred from Ratios r1








LIT 0.2009856574 27 124 Up regulated by BR late (between 3 and 24

hours) (Table S 1 PubmedID:15681342)

LIT 0.2033985714 14 63 Arabidopsis genomic regions down-regulated ex-

pression upon rrp4est mutation (TableS2 Pub-

medID:18160042)

LIT 0.2037863636 2 3 dn,Differentially expressed genes in jai3-1 vs

wild-type plants , after JA treatment

40


https://doi.org/10.1101/023051

LIT 0.2044906693 33 176 Repressed in the vfb mutants compared to the

wild type (Table S2 PubmedID:17435085)

LIT 0.2057408627 246 1466 Induced in PPV-infected Arabidopsis leaf tis-

sues 17 days post inoculation (Table S1 Pub-

medID:18613973)

LIT 0.2134856202 126 730 Down-6x2-List of genes that are upregulated or


PubmedID:20406451)

LIT 0.2134856202 70 415 Significant up-regulation in 72hrs in re-

sponse to iron-deprivation (Table S2 Pubme-

dID:18436742)

LIT 0.2134856202 112 631 down-Genes with altered expression above 1.3

lower bound fold change in bol-D young leaves

LIT 0.2138952672 5 15 dn,aox1a normal vs Col-0 normal,Relative tran-



bidopsis

LIT 0.2138952672 41 223 Repressed at Least 1.5-Fold in the Crab Shell

Chitin Mixture-Treated Seedlings (Table S 5

PubmedID:15923325)

LIT 0.2138952672 59 310 Down-regulated genes that were nonadditively

expressed in allopolyploid line 1200 relative to

the 1:1 parent mix–parentmix 5250 FDR Com-

monVar (Dataset S5 PubmedID:19274085)

LIT 0.2138952672 41 210 Down-regulated (LU) Gene List-Cold Late (Ta-

ble S 9 PubmedID:16214899)

LIT 0.2139394361 16 72 Up in 2-6, bound (by fdr) (Table S3 Pubme-

dID:20675573)

LIT 0.2139394361 7 35 Aza-dC downregulated (TableS1 Pubme-

dID:15516340)

LIT 0.2139394361 9 32 Strong upregulation of transcripts from infected

roots of plants –clubroot disease

LIT 0.2139394361 7 21 up,gl3-3-trichome,The 25 most up- and down-

regulated genes in gl3-3 and try-JC trichomes

LIT 0.2145473606 42 221 genes that are induced in Ler and hy5-1:

LIT 0.2145473606 90 481 Up-regulated genes that were nonadditively ex-

pressed in allopolyploid line 1200 relative to

the 1:1 parent mix–parentmix 1250 FDR Per-

GeneVar (Dataset S3 PubmedID:19274085)

41


https://doi.org/10.1101/023051

LIT 0.2145473606 8 35 All Genes That Pass Significance Criteria fol-

lowing Induction of GLK2.

42


https://doi.org/10.1101/023051

References29

DevelopmentCoreTeam (2014) R: A Language and Environment for Statistical Computing. Vi-30

enna.31

43


https://doi.org/10.1101/023051

phenotypic and genomic di erentiation of arabidopsis thaliana...2015/07/22 · 68 swiss alps (luo...

Documents